Decoding method and apparatus, encoding method and apparatus, and corresponding encoder and decoder

Abstract

The present application relates to the technical field of encoding and decoding, and provides a decoding method and apparatus, an encoding method and apparatus, and a corresponding encoder and a corresponding decoder. The decoding method comprises: acquiring a bitstream of an image to be decoded, the image to be decoded comprising a block to be decoded; then, acquiring a reconstructed low-frequency sub-band of a reference image corresponding to the image to be decoded, the reconstructed low-frequency sub-band being obtained on the basis of wavelet coefficients of a low-frequency sub-band obtained after the reference image is subjected to wavelet transform; next, on the basis of the reconstructed low-frequency sub-band, performing inter prediction on the block to be decoded to obtain prediction data of the block to be decoded; and finally, on the basis of the prediction data, decoding the bitstream to obtain reconstructed data of the block to be decoded. The method can improve the encoding and decoding efficiency.

Description

Decoding methods, encoding methods and devices, and corresponding encoders and decoders

[0001] This application claims priority to Chinese Patent Application No. 202510123226.9, filed on January 24, 2025, entitled "Decoding Method, Encoding Method and Apparatus, Corresponding Encoder and Decoder", filed on July 8, 2025, filed on the same date, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of image processing technology, and in particular to a decoding method, encoding method and apparatus, and corresponding encoder and decoder. Background Technology

[0003] Digital video capabilities can be incorporated into a wide variety of devices, including digital television, digital live broadcasting systems, wireless broadcasting systems, personal digital assistants (PDAs), laptops or desktop computers, tablets, e-book readers, digital cameras, digital recording devices, digital media players, video game devices, video game consoles, cellular or satellite radio phones (so-called "smartphones"), video conferencing devices, video streaming devices, and the like. Digital video devices implement video compression techniques, such as those described in standards defined by MPEG-2, MPEG-4, ITU-T H.263, ITU-T H.264 / MPEG-4 Part 10 High-Level Video Coding (AVC), the H.265 / HEVC video coding standard, and extensions to such standards. By implementing such video compression techniques, video devices can transmit, receive, encode, decode, and / or store digital video information more efficiently.

[0004] Video compression techniques perform spatial (inter-image) prediction and / or temporal (inter-image) prediction to reduce or remove redundancy inherent in video sequences. For block-based video coding, video stripes (i.e., video frames or portions of video frames) can be divided into several image blocks, which may also be referred to as tree blocks, coding units (CUs), and / or coding nodes.

[0005] However, the current encoding and decoding methods have low efficiency and need to be improved. Summary of the Invention

[0006] To address the aforementioned technical problems, this application provides a decoding method, an encoding method and apparatus, and corresponding encoders and decoders, which can improve encoding and decoding efficiency.

[0007] In a first aspect, embodiments of this application provide a decoding method, the method comprising: acquiring a bitstream of an image to be decoded, the image to be decoded including blocks to be decoded; acquiring a reconstructed low-frequency subband of a reference image corresponding to the image to be decoded, the reconstructed low-frequency subband being obtained based on wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image; performing inter-frame prediction on the blocks to be decoded based on the reconstructed low-frequency subband to obtain prediction data of the blocks to be decoded; and decoding the bitstream based on the prediction data to obtain reconstructed data of the blocks to be decoded.

[0008] In this embodiment, instead of obtaining a reference frame for inter-frame prediction of the current block within the current frame, this embodiment uses the reconstructed low-frequency subband data (i.e., reconstructed low-frequency subband) of the reference frame to perform inter-frame prediction of the current block within the current frame. The size of the low-frequency subband of the reference frame is one-quarter of the reference frame's size, which significantly reduces the amount of reference data obtained during inter-frame prediction, thereby reducing the bit width occupied by the reference data. Since it is not necessary to read the entire image data of the reference frame, but only its reconstructed low-frequency subband, memory usage and bandwidth consumption can be reduced. Furthermore, using the reconstructed low-frequency subband of the reference frame to perform inter-frame prediction of the current block to obtain the predicted value of the current block for decoding the current frame improves the efficiency of inter-frame prediction compared to using the reference frame directly, thus improving decoding efficiency.

[0009] Based on the first aspect, in some possible implementations, the block to be decoded is the block data in the wavelet coefficients of the low-frequency subband obtained by wavelet transform of the image to be decoded during encoding.

[0010] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is only 1 / 4 of the image before the wavelet transform (e.g., the image to be decoded), obtaining the reference block based on the reconstructed low-frequency subband reduces the amount of data acquired compared to obtaining the reference block based on the reference frame, thereby improving inter-frame prediction efficiency and decoding efficiency. Furthermore, considering that the data distribution of the reconstructed low-frequency subband is more uniform than that of the reconstructed high-frequency subband output by the wavelet transform, determining the data from the reconstructed low-frequency subband as the prediction data for the block to be decoded results in more accurate prediction data, improving the accuracy of inter-frame prediction and thus enhancing decoding accuracy.

[0011] Based on the first aspect, in some possible implementations, the prediction data is data from a reference block obtained from reconstructing the low-frequency subband.

[0012] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is 1 / 4 of the image before the wavelet transform (e.g., the image to be decoded), obtaining the reference block based on the reconstructed low-frequency subband reduces the amount of data acquired compared to obtaining the reference block based on the reference frame, thereby improving inter-frame prediction efficiency and coding efficiency. Considering that the data distribution of the reconstructed low-frequency subband is more uniform than that of the reconstructed high-frequency subband output by the wavelet transform, determining the data of the reconstructed low-frequency subband as the prediction data for the block to be coded results in more accurate prediction data, improving the accuracy of inter-frame prediction and thus improving coding accuracy.

[0013] Based on the first aspect, in some possible implementations, the prediction data is data of a reference block obtained from a first region of the reconstructed low-frequency subband, the first region being a region obtained by expanding a specified number of samples around the corresponding position of the block to be decoded in the reconstructed low-frequency subband.

[0014] In this embodiment, when performing inter-frame prediction on the current block within the current frame based on the reconstructed low-frequency subband of the reference frame, the reference block data can be obtained in a first region of the reconstructed low-frequency subband as prediction data for the current block's inter-frame prediction. This ensures that the reference block is not far from the current block, but rather within the controllable first region, thereby improving the efficiency of obtaining the prediction block, thus improving the prediction efficiency of inter-frame prediction, and reducing the amount of decoding data to improve decoding efficiency. Furthermore, the first region is obtained by expanding the area around the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame, ensuring that this region is still near the corresponding position of the current block. This means that the reference block obtained from inter-frame prediction is also quite similar to the current block, ensuring the accuracy of inter-frame prediction, thereby reducing the amount of decoding data and improving decoding efficiency. In addition, performing inter-frame prediction within the first region to obtain the prediction data for the current block reduces decoding complexity compared to determining the prediction data for the current block within the entire reference frame.

[0015] Based on the first aspect, in some possible implementations, the first region is a region obtained by expanding the corresponding position of the block to be decoded horizontally to both sides by the same number of first points.

[0016] In this embodiment of the application, since the first region is the region used to determine the position of the reference block for inter-frame prediction of the current block, the first region can be set as the region obtained by expanding the same number of samples to the left and right sides in the horizontal direction with the corresponding position of the current block in the reconstructed low-frequency sub-band as the center.

[0017] Based on the first aspect, in some possible implementations, the first region is a region obtained by expanding the corresponding position of the block to be decoded by the same number of second sample points in the vertical direction to both sides.

[0018] In this embodiment of the application, since the first region is the region used to determine the position of the reference block for inter-frame prediction of the current block, the first region can be set as the region obtained by expanding the same number of samples to the left and right sides in the vertical direction with the corresponding position of the current block in the reconstructed low-frequency sub-band as the center.

[0019] Based on the first aspect, in some possible implementations, the number of first sample points is greater than the number of second sample points.

[0020] In this embodiment, the number of samples expanded horizontally is greater than the number of samples expanded vertically. In this embodiment, considering that the decoder reads data in the reconstructed low-frequency subband of the reference frame line by line, if the number of samples is greater than the number of samples, the decoder will read more lines of data in the first region of the low-frequency subband during line-by-line reading, resulting in excessive memory consumption. Therefore, in this embodiment, the specified number of samples expanded horizontally in the first region is greater than the specified number of samples expanded vertically, which reduces the number of lines of data read by the decoder in the first region, thereby reducing memory consumption during inter-frame prediction.

[0021] Based on the first aspect, in some possible implementations, the size of the block to be decoded is 8x8, the number of first samples is 7, and the number of second samples is 3.

[0022] In this embodiment, the block size is 8x8, and it is expanded by 7 samples on both sides in the horizontal direction and by 3 samples on both sides in the vertical direction to obtain a first region. The first region obtained in this way can be the region with the highest probability of occurrence of the reference block for the inter-frame prediction of the current block. Therefore, performing inter-frame prediction in this first region to obtain the predicted block of the current block can improve the prediction efficiency and accuracy of the current block, thereby improving the decoding speed and decoding efficiency.

[0023] Secondly, embodiments of this application provide an encoding method, which includes: obtaining a reconstructed low-frequency subband of a reference image corresponding to an image to be encoded, wherein the reconstructed low-frequency subband is obtained based on wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image, and the image to be encoded includes a block to be encoded; performing inter-frame prediction on the block to be encoded based on the reconstructed low-frequency subband to obtain prediction data of the block to be encoded; and encoding the block to be encoded based on the prediction data to obtain a bitstream.

[0024] In this embodiment, instead of obtaining a reference frame for inter-frame prediction of the current block within the current frame, this embodiment uses the reconstructed low-frequency subband data (i.e., reconstructed low-frequency subband) of the reference frame to perform inter-frame prediction of the current block within the current frame. The size of the low-frequency subband of the reference frame is one-quarter of the reference frame's size, which significantly reduces the amount of reference data obtained during inter-frame prediction, thereby reducing the bit width occupied by the reference data. Since it is not necessary to read the entire image data of the reference frame, but only its reconstructed low-frequency subband, memory usage and bandwidth consumption can be reduced. Furthermore, using the reconstructed low-frequency subband of the reference frame to perform inter-frame prediction of the current block to obtain the predicted value of the current block for decoding the current frame improves the efficiency of inter-frame prediction compared to using the reference frame directly, thus improving decoding efficiency.

[0025] Based on the second aspect, in some possible implementations, the prediction data is data from a reference block obtained from reconstructing the low-frequency subband.

[0026] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is 1 / 4 of the image before the wavelet transform (e.g., the image to be decoded), obtaining the reference block based on the reconstructed low-frequency subband reduces the amount of data acquired compared to obtaining the reference block based on the reference frame, thereby improving inter-frame prediction efficiency and coding efficiency. Considering that the data distribution of the reconstructed low-frequency subband is more uniform than that of the reconstructed high-frequency subband output by the wavelet transform, determining the data of the reconstructed low-frequency subband as the prediction data for the block to be coded results in more accurate prediction data, improving the accuracy of inter-frame prediction and thus improving coding accuracy.

[0027] Based on the second aspect, in some possible implementations, the prediction data is data of a reference block obtained from a first region of the reconstructed low-frequency subband, the first region being a region obtained by expanding a specified number of samples around the corresponding position of the block to be encoded in the reconstructed low-frequency subband.

[0028] In this embodiment, when performing inter-frame prediction on the current block within the current frame based on the reconstructed low-frequency subband of the reference frame, the data of the reference block can be obtained in a first region within the reconstructed low-frequency subband as prediction data for the current block's inter-frame prediction. Compared to searching for a reference block similar to the current block within the reconstructed low-frequency subband of the reference frame, this embodiment searches within a first region, which narrows the range of reference blocks, ensuring that the reference blocks are not far from the current block and remain within the controllable first region. This reduces the bit width of the current block's MV, thereby improving the prediction efficiency of inter-frame prediction and reducing the amount of encoded data to improve coding efficiency. Furthermore, the first region is obtained by expanding the area around the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame, ensuring that the region is still near the corresponding position of the current block. This ensures that the reference blocks searched for in inter-frame prediction are also similar to the current block, guaranteeing the accuracy of inter-frame prediction, thereby reducing the amount of encoded data and improving coding efficiency. In this embodiment of the application, inter-frame prediction is performed in the first region to obtain the prediction data of the current block. Compared with determining the prediction data of the current block in the entire reference frame, the range of the MV of the current block can be narrowed, so that the MV of the current block is fixed within a certain range.

[0029] Based on the second aspect, in some possible implementations, the first region is a region obtained by expanding the corresponding position of the block to be encoded by the same number of first points in the horizontal direction.

[0030] In this embodiment, since the first region is the search range for searching the MV of the current block, it is set as a region centered on the corresponding position of the current block in the reconstructed low-frequency subband, expanding the same number of samples to the left and right sides in the horizontal direction. This ensures the accuracy of the searched MVs in scenarios where the MV of the current block is searched with equal probability on both sides of the corresponding position in the horizontal direction.

[0031] Based on the second aspect, in some possible implementations, the first region is a region obtained by expanding the corresponding position of the block to be encoded by the same number of second sample points in the vertical direction to both sides.

[0032] In this embodiment, since the first region is the search range for searching the MV of the current block, it is set as a region centered on the corresponding position of the current block in the reconstructed low-frequency subband, and expanded vertically to both the upper and lower sides with the same number of samples. This ensures the accuracy of the searched MVs in scenarios where the MV of the current block is searched with equal probability on both sides of the corresponding position in the vertical direction.

[0033] Based on the second aspect, in some possible implementations, the number of first sample points is greater than the number of second sample points.

[0034] Thus, the number of samples expanded horizontally is greater than the number of samples expanded vertically. In this embodiment, considering that the encoder reads data in the reconstructed low-frequency subband of the reference frame line by line from the hardware, if the number of samples is greater than the number of samples, the encoder will read fewer lines of data when reconstructing the data in the first region of the low-frequency subband. Therefore, the number of times line data needs to be read increases, requiring more line buffers and thus consuming too much memory. Therefore, in this embodiment, the specified number of samples expanded horizontally in the first region is greater than the specified number of samples expanded vertically. This reduces the search range of the MV while also reducing the number of times the encoder reads line data in the first region, thereby reducing memory consumption during inter-frame prediction.

[0035] Based on the second aspect, in some possible implementations, the size of the block to be encoded is 8x8, the number of first sample points is 7, and the number of second sample points is 3.

[0036] In this embodiment, the block size is 8x8, and it is expanded by 7 samples on both sides in the horizontal direction and by 3 samples on both sides in the vertical direction to obtain a first region. The first region obtained in this way can be the region with the highest probability of occurrence of the reference block for the inter-frame prediction of the current block. Therefore, performing inter-frame prediction in this first region to obtain the predicted block of the current block can improve the prediction efficiency and accuracy of the current block, thereby improving the coding speed and coding efficiency.

[0037] Based on the second aspect, in some possible implementations, the block to be encoded is the block data in the wavelet coefficients of the low-frequency subband obtained after wavelet transform of the image to be encoded.

[0038] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is 1 / 4 of the image before the wavelet transform (e.g., the image to be encoded), performing inter-frame prediction on the block data in the low-frequency subband reduces the amount of data encoded compared to performing inter-frame prediction and encoding on the image to be encoded, thereby improving the efficiency of inter-frame prediction and encoding. Considering that the data distribution in the low-frequency subband is more uniform than that in the high-frequency subband output by the wavelet transform, performing inter-frame prediction and encoding on the block data in the low-frequency subband using the method of this embodiment can improve the accuracy of inter-frame prediction, thereby improving the encoding accuracy. Furthermore, it allows the encoded block data of the low-frequency subband to be used for display, for example, the displayed image of the low-frequency subband can be a thumbnail of the image to be encoded (e.g., the original image to be encoded).

[0039] Thirdly, embodiments of this application provide a decoding apparatus, which may include: an acquisition module for acquiring a bitstream of an image to be decoded, the image to be decoded including a block to be decoded; the acquisition module is further configured to acquire a reconstructed low-frequency subband of a reference image corresponding to the image to be decoded, the reconstructed low-frequency subband being obtained based on wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image; a prediction module for performing inter-frame prediction on the block to be decoded based on the reconstructed low-frequency subband to obtain prediction data of the block to be decoded; and a decoding module for decoding the bitstream based on the prediction data to obtain reconstructed data of the block to be decoded.

[0040] Fourthly, embodiments of this application provide an encoding apparatus, which includes: an acquisition module, configured to acquire a reconstructed low-frequency subband of a reference image corresponding to an image to be encoded, wherein the reconstructed low-frequency subband is obtained based on wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image, and the image to be encoded includes a block to be encoded; a prediction module, configured to perform inter-frame prediction on the block to be encoded based on the reconstructed low-frequency subband to obtain prediction data for the block to be encoded; and an encoding module, configured to encode the block to be encoded based on the prediction data to obtain a bitstream.

[0041] Fifthly, a system for distributing bitstreams is provided, the system comprising: at least one storage medium for storing a bitstream generated according to the method of the first or second aspect and any embodiment thereof; and a streaming media device for acquiring the bitstream from the at least one storage medium and transmitting the bitstream, wherein the streaming media device includes a content server or a content distribution server.

[0042] In a sixth aspect, a transcoding system is provided, comprising: at least one storage medium for storing a bitstream generated according to the method of the second aspect or any embodiment thereof; and a transcoding device for acquiring the bitstream from the at least one storage medium and transcoding the bitstream.

[0043] In one embodiment, the transcoding device can convert the bitstream into MPEG-4 Part 4 (MP4), Matroska Video (MKV), Audio Video Interleave (AVI), Digital Audio Video (DAV), etc., without limitation.

[0044] In a seventh aspect, this application provides an apparatus comprising: one or more processors; a memory for storing one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in the first aspect or any embodiment of the first aspect.

[0045] Eighthly, this application provides an apparatus comprising: one or more processors; a memory for storing one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in the second aspect or any embodiment of the second aspect above.

[0046] Ninthly, this application provides a computer-readable storage medium including a computer program that, when executed on a device, causes the device to perform the method described in the first aspect or any embodiment of the first aspect.

[0047] In a tenth aspect, this application provides a computer-readable storage medium including a computer program that, when executed on a device, causes the device to perform the method described in the second aspect or any of the embodiments of the second aspect.

[0048] In one aspect, this application provides a computer program that, when executed by a device, is used to perform the method described in the first aspect or any embodiment of the first aspect, or to perform the method described in the second aspect or any embodiment of the second aspect.

[0049] In a twelfth aspect, this application provides a computer program product comprising computer program code that, when executed on a device, causes the device to perform the method described in the first aspect or any embodiment of the first aspect, or to perform the method described in the second aspect or any embodiment of the second aspect.

[0050] In a thirteenth aspect, this application provides a bitstream generated according to the method in the second aspect or any embodiment of the second aspect described above.

[0051] In a fourteenth aspect, this application provides a computer-readable storage medium storing a bitstream generated according to the method in the second aspect or any embodiment of the second aspect described above.

[0052] In a fifteenth aspect, an apparatus for storing a bitstream is provided, the apparatus comprising: a transceiver unit and a storage unit, the transceiver unit being configured to receive a bitstream generated according to the method described in accordance with the second aspect or any embodiment thereof, and the storage unit being configured to store the bitstream.

[0053] In a sixteenth aspect, an apparatus for transmitting a bitstream is provided, the apparatus comprising: a storage unit and a transceiver unit, the storage unit being configured to store a bitstream generated according to the method described in accordance with the second aspect or any embodiment thereof, and the transceiver unit being configured to transmit the bitstream.

[0054] In a seventeenth aspect, an encoder is provided. The encoder includes a processing circuit that implements the steps described in the second aspect or any embodiment of the second aspect.

[0055] In an eighteenth aspect, a decoder is provided. The decoder includes processing circuitry that implements the steps described in the first aspect or any embodiment of the first aspect. Attached Figure Description

[0056] Figure 1A is a schematic block diagram of a video encoding and decoding system provided in an embodiment of this application;

[0057] Figure 1B is a schematic block diagram of a video decoding system provided in an embodiment of this application;

[0058] Figure 2 is a schematic block diagram of an encoder provided in an embodiment of this application;

[0059] Figure 3 is a schematic block diagram of a decoder provided in an embodiment of this application;

[0060] Figure 4 is a schematic diagram of the structure of a video encoding device provided in an embodiment of this application;

[0061] Figure 5 is a schematic diagram of the structure of a device provided in an embodiment of this application;

[0062] Figure 6 is a schematic block diagram of an encoder based on wavelet transform provided in an embodiment of this application;

[0063] Figure 7 is a schematic block diagram of a wavelet transform-based decoder provided in an embodiment of this application;

[0064] Figure 8 is a schematic block diagram of an encoder provided in an embodiment of this application;

[0065] Figure 9A is a schematic diagram of sub-graph division provided in an embodiment of this application;

[0066] Figure 9B is a schematic diagram of sub-graph partitioning provided in an embodiment of this application;

[0067] Figure 10 is a schematic diagram of wavelet transform provided in an embodiment of this application;

[0068] Figure 11 is a schematic block diagram of a decoder provided in an embodiment of this application;

[0069] Figure 12 is a schematic block diagram of an encoder provided in an embodiment of this application;

[0070] Figure 13A is a schematic diagram of the structure of an image bitstream provided in an embodiment of this application;

[0071] Figure 13B is a schematic diagram of the structure of an image bitstream provided in an embodiment of this application;

[0072] Figure 14A is a schematic diagram of the structure of an image bitstream provided in an embodiment of this application;

[0073] Figure 14B is a schematic diagram of the structure of an image bitstream provided in an embodiment of this application;

[0074] Figure 15 is a schematic diagram of the structure of an image bitstream provided in an embodiment of this application;

[0075] Figures 16A to 16C are schematic diagrams of the structure of an image bitstream provided in an embodiment of this application;

[0076] Figures 17A and 17B are schematic diagrams of the structure of an image bitstream provided in an embodiment of this application;

[0077] Figure 18 is a schematic block diagram of a decoder provided in an embodiment of this application;

[0078] Figure 19 is a schematic block diagram of a decoder provided in an embodiment of this application;

[0079] Figure 20a is a schematic diagram of an edge-cloud system provided in an embodiment of this application;

[0080] Figure 20b is a schematic diagram of an encoding method provided in an embodiment of this application;

[0081] Figure 21a is a schematic diagram of an encoding method provided in an embodiment of this application;

[0082] Figure 21b is a schematic diagram of an encoding method provided in an embodiment of this application;

[0083] Figure 21c is a schematic diagram of expanding the reconstructed low-frequency subband according to an embodiment of this application;

[0084] Figure 22 is a schematic diagram of an encoding method provided in an embodiment of this application;

[0085] Figure 23 is a schematic diagram of a decoding method provided in an embodiment of this application;

[0086] Figure 24 is a schematic diagram of a decoding method provided in an embodiment of this application;

[0087] Figure 25a is a schematic diagram of a decoding device provided in an embodiment of this application;

[0088] Figure 25b is a schematic diagram of an encoding device provided in an embodiment of this application. Detailed Implementation

[0089] The embodiments of this application are described below with reference to the accompanying drawings. In the following description, reference is made to the accompanying drawings, which form part of this application and illustrate specific aspects of the embodiments of this application or to which specific aspects of the embodiments of this application may be used. It should be understood that the embodiments of this application may be used in other aspects and may include structural or logical variations not depicted in the drawings. Therefore, the following detailed description should not be construed in a limiting sense, and the scope of this application is defined by the appended claims. For example, it should be understood that the disclosure of the described methods is equally applicable to corresponding devices or systems for performing the methods, and vice versa. For example, if one or more specific method steps are described, the corresponding device may include one or more units, such as functional units, to perform the described one or more method steps (e.g., one unit performs one or more steps, or multiple units, each performing one or more of multiple steps), even if such one or more units are not explicitly described or illustrated in the drawings. On the other hand, for example, if a specific apparatus is described based on one or more units such as functional units, the corresponding method may include a step to perform the functionality of one or more units (e.g., a step to perform the functionality of one or more units, or multiple steps, each of which performs the functionality of one or more units among a plurality of units), even if such one or more steps are not explicitly described or illustrated in the accompanying drawings. Furthermore, it should be understood that, unless otherwise expressly stated, features of the various exemplary embodiments and / or aspects described herein can be combined with each other.

[0090] The terminology used in the implementation section of this application is for the purpose of explaining specific embodiments of this application only, and is not intended to limit this application.

[0091] The following is a brief introduction to some concepts that may be involved in the embodiments of this application. These concepts are only used to explain the specific embodiments of this application and are not intended to limit this application.

[0092] The residual is the difference between the reconstructed value (or actual value) of a sample or data element and its predicted value.

[0093] A residual block is an M×N residual matrix composed of the residuals corresponding to the coded blocks.

[0094] Dequantization is the process of scaling the quantized residual to obtain the reconstructed residual value.

[0095] A partition divides a set into subsets. Each element in the set belongs to one and only one subset.

[0096] Partition type: The way the subsets obtained from the partition are organized.

[0097] A decoded picture is an image reconstructed by the decoder based on the bitstream.

[0098] Prediction is the specific implementation of the prediction process.

[0099] The prediction process uses previously decoded samples to obtain the predicted value for the current sample.

[0100] Syntax element: The result of parsing data units in a bitstream.

[0101] A bitstream is a binary data stream that encodes all or part of an image sample.

[0102] Video coding generally refers to the processing of a sequence of images that form a video or video sequence. In the field of video coding, the terms "picture," "frame," or "image" can be used synonymously. Video coding is performed on the source side and typically involves processing (e.g., by compression) the raw video images to reduce the amount of data required to represent them, thus enabling more efficient storage and / or transmission. Video decoding is performed on the destination side and typically involves inverse processing relative to the encoder to reconstruct the video images. The combination of encoding and decoding is also known as encoding and decoding.

[0103] A video sequence consists of a series of pictures, which are further divided into slices, and slices into blocks. Video coding is performed on a block-by-block basis. In some newer video coding standards, the concept of a block has been further expanded. For example, the H.264 standard uses macroblocks (MBs), which can be further divided into multiple prediction blocks (partitions) for predictive coding. The High Efficiency Video Coding (HEVC) standard uses basic concepts such as coding units (CUs), prediction units (PUs), and transform units (TUs) to functionally divide various block units, and employs a novel tree-based structure for description. For instance, a CU can be divided into smaller CUs using a quadtree, and these smaller CUs can be further divided, forming a quadtree structure. The CU is the basic unit for partitioning and encoding the image. Similar tree structures exist for PUs and TUs. A PU corresponds to a prediction block and is the basic unit for predictive coding. CUs are further divided into multiple PUs according to partitioning patterns. TU can correspond to a transform block, which is the basic unit for transforming the prediction residual. However, regardless of CU, PU, ​​or TU, they all essentially belong to the concept of block data (or image blocks).

[0104] For example, in HEVC, the CTU is split into multiple CUs using a quadtree structure represented as a coding tree. At the CU level, a decision is made on whether to use inter-picture (temporal) or intra-picture (spatial) prediction to encode picture regions. Each CU can be further split into one, two, or four PUs based on the PU splitting type. The same prediction process is applied within a PU, and relevant information is transmitted to the decoder based on the PU. After obtaining residual blocks by applying the prediction process based on the PU splitting type, the CU can be segmented into transform units (TUs) according to other quadtree structures similar to the coding tree used for CUs. In the latest developments in video compression technology, quadtree and binary tree (QTBT) frame segmentation is used to divide coding blocks. In the QTBT block structure, CUs can be square or rectangular in shape.

[0105] In this paper, for ease of description and understanding, the block of data to be processed in the current image is referred to as the current block. For example, in encoding, the current block refers to the block currently being encoded; in decoding, the current block refers to the block currently being decoded. The decoded current block in the reference image used to predict the current block is called the reference block. That is, the reference block is the block that provides a reference signal for the current block, where the reference signal represents the pixel value within the image block. In the inter-frame prediction process, the block in the reference image that provides the prediction signal for the current block is called the prediction block, where the prediction signal represents the pixel value, sampled value, or sampled signal within the prediction block. For example, after traversing multiple reference blocks, an optimal reference block is found. This optimal reference block will provide prediction for the current block; this block is called the prediction block.

[0106] In lossless video coding, the original video image can be reconstructed, meaning the reconstructed video image has the same quality as the original (assuming no transmission loss or other data loss during storage or transmission). In lossy video coding, further compression is performed, for example, through quantization, to reduce the amount of data required to represent the video image. However, the decoder cannot fully reconstruct the video image, meaning the quality of the reconstructed video image is lower or worse than the original video image.

[0107] YUV is a color space model used to represent colors, widely used in image encoding and decoding, video encoding and decoding, digital image processing, television broadcasting, and other fields. YUV separates the luminance information from the chrominance information of an image.

[0108] The three components of YUV:

[0109] Y (luminance) component: Y represents the luminance information of an image, that is, the brightness or darkness of the image. It is obtained by weighting the red, green, and blue color channels according to certain weights. The Y component plays an important role in the sharpness and detail of an image.

[0110] U (chromaticity) component: U represents the chromaticity information of the image, indicating the offset of the blue channel relative to the luminance Y. It measures the change in the blue component.

[0111] V (chromaticity) component: V represents the chromaticity information of the image, indicating the offset of the red channel relative to the luminance Y. Residual: The difference between the reconstructed value (or actual value) of a sample or data element and its predicted value.

[0112] The encoding / decoding method of this application embodiment encodes and decodes images or videos on a block-by-block basis.

[0113] In some embodiments, the block to be encoded or the block to be decoded may be an image block or a video block obtained from an image or video.

[0114] In some embodiments, the block to be encoded or the block to be decoded may be an image block or video block within a subgraph obtained by dividing the image or video into subgraphs.

[0115] In some embodiments, the block to be encoded or the block to be decoded can be a block obtained from the transformed data of the image or video to be encoded after transformation processing (e.g., wavelet transform (also known as wavelet forward transform, etc., without limitation).

[0116] This application does not restrict the specific division method of the blocks to be encoded or decoded in an image or video, nor does it restrict the method of obtaining the blocks.

[0117] On the encoding side, the image to be encoded can be a frame from an image or video, a sub-image obtained by dividing a frame from an image or video, an image obtained after transformation (such as wavelet transform), or a sub-image from the transformed image; there are no restrictions here. On the encoding side, "current frame" can represent "image to be encoded," and "current block" can represent "block to be encoded." A frame is a frame of image to be displayed.

[0118] Similarly, the decoding side corresponds to the encoding side. On the decoding side, the image to be decoded can be a frame in an image or video, a sub-image obtained by dividing a frame in an image or video, an image obtained after transformation (such as wavelet transform), or a sub-image in the transformed image; there are no restrictions here. On the decoding side, "current frame" can represent "image to be decoded," and "current block" can represent "block to be decoded." A frame is a frame of image to be displayed.

[0119] Whether on the encoding or decoding side, the reconstructed data (also called the reconstructed value) of a block can be described as a reconstructed block, and the already encoded or decoded block referenced when encoding or decoding the current block can be described as a "reference block".

[0120] Whether on the encoding or decoding side, the reference image, also known as the reference frame, can be a frame image encoded before the current frame in the video sequence, or, when the current frame is a sub-frame, the reference image is a sub-frame encoded before the current frame, without any restrictions.

[0121] Whether on the encoding or decoding side, the prediction data for the current block is also referred to as the prediction block or the prediction value for the current block.

[0122] Whether on the encoding or decoding side, reconstruction is also referred to as refactoring.

[0123] Whether on the encoding or decoding side, the bitstream is also described as a bitstream, etc.

[0124] Whether on the encoding or decoding side, the residual is also referred to as a residual block.

[0125] Whether on the encoding or decoding side, the actual value of the current block is also expressed as the original value of the current block, or the value to be encoded in the current block.

[0126] The dashed boxes or dashed arrows in the accompanying drawings of the various embodiments of this application indicate that the step is optional, and the data transmitted indicated by the arrow is optional.

[0127] The system architecture used in the embodiments of this application is described below. Referring to FIG1A, FIG1A provides an exemplary schematic block diagram of a video encoding and decoding system 10 used in the embodiments of this application. As shown in FIG1A, the video encoding and decoding system 10 may include a source device 12 and a destination device 14. The source device 12 generates encoded video data, and therefore, the source device 12 may be referred to as a video encoding device. The destination device 14 can decode the encoded video data generated by the source device 12, and therefore, the destination device 14 may be referred to as a video decoding device. Various embodiments of the source device 12, the destination device 14, or both may include one or more processors and memory coupled to the one or more processors. The memory may include, but is not limited to, RAM, ROM, EEPROM, flash memory, or any other media that can be used to store desired program code in the form of computer-accessible instructions or data structures. The source device 12 and the destination device 14 may include a variety of devices, including desktop computers, mobile computing devices, notebook (e.g., laptop) computers, tablet computers, set-top boxes, handsets, televisions, cameras, display devices, digital media players, video game consoles, in-vehicle computers, wireless communication devices, or the like.

[0128] Source device 12 and destination device 14 can communicate via link 13, through which destination device 14 can receive encoded video data from source device 12. Link 13 may include one or more media or devices capable of transmitting encoded video data from source device 12 to destination device 14. In one example, link 13 may include one or more communication media enabling source device 12 to transmit encoded video data to destination device 14 in real time. In this example, source device 12 may modulate the encoded video data according to a communication standard (e.g., a wireless communication protocol) and transmit the modulated video data to destination device 14. The one or more communication media may include wireless and / or wired communication media, such as radio frequency (RF) spectrum or one or more physical transmission lines. The one or more communication media may form part of a packet-based network, such as a local area network, wide area network, or global network (e.g., the Internet). The one or more communication media may include routers, switches, base stations, or other devices facilitating communication from source device 12 to destination device 14.

[0129] The source device 12 includes an encoder 20. Optionally, the source device 12 may also include an image source 16, an image preprocessor 18, and a communication interface 22. In specific implementations, the encoder 20, image source 16, image preprocessor 18, and communication interface 22 may be hardware components or software programs within the source device 12. These are described below:

[0130] Image source 16 may include or be any type of image capture device for, for example, capturing real-world images, and / or any type of image or commentary (for screen content encoding, some text on the screen is also considered as an image to be encoded or part of an image) generation device, such as a computer graphics processor for generating computer-animated images, or any type of device for acquiring and / or providing real-world images, computer-animated images (e.g., screen content, virtual reality (VR) images), and / or any combination thereof (e.g., augmented reality (AR) images). Image source 16 may be a camera for capturing images or a memory for storing images. Image source 16 may also include any type of (internal or external) interface for storing previously captured or generated images and / or acquiring or receiving images. When image source 16 is a camera, image source 16 may be, for example, a local or integrated camera integrated into a source device; when image source 16 is a memory, image source 16 may be a local or integrated memory integrated into a source device. When the image source 16 includes an interface, the interface may be, for example, an external interface for receiving images from an external video source, such as an external image capture device, like a camera, external storage, or an external image generation device, such as an external computer graphics processor, computer, or server. The interface can be any type of interface according to any proprietary or standardized interface protocol, such as a wired or wireless interface, or an optical interface.

[0131] An image can be viewed as a two-dimensional array or matrix of pixels. Pixels in the array are also called sampling points. The number of sampling points in the array or image along the horizontal and vertical directions (or axes) defines the image's size and / or resolution. To represent color, three color components are typically used; that is, an image can be represented as or contain three sampling arrays. For example, in RBG format or color space, an image includes corresponding red, green, and blue sampling arrays. However, in video coding, each pixel is typically represented in a luma / chroma format or color space. For example, for a YUV format image, this includes a luma component indicated by Y (sometimes also indicated by L) and two chroma components indicated by U and V. The luma component Y represents the brightness or grayscale level intensity (e.g., both are the same in a grayscale image), while the two chroma components U and V represent chroma or color information components. Accordingly, a YUV format image includes a luma sampling array of luma sample values ​​(Y) and two chroma sampling arrays of chroma values ​​(U and V). An RGB format image can be converted or transformed to YUV format, and vice versa; this process is also called color transformation or conversion. If the image is black and white, it may only include a luminance sampling array. In this embodiment, the image transmitted from image source 16 to image processor can also be referred to as raw image data 17.

[0132] Image preprocessor 18 is configured to receive raw image data 17 and perform preprocessing on the raw image data 17 to obtain a preprocessed image 19 or preprocessed image data 19. For example, the preprocessing performed by image preprocessor 18 may include retouching, color format conversion (e.g., from RGB format to YUV format), color correction, or noise reduction.

[0133] Encoder 20 (or video encoder 20) is used to receive preprocessed image data 19 and process the preprocessed image data 19 using a relevant prediction mode (such as the prediction mode in the various embodiments herein) to provide encoded image data 21. In some embodiments, encoder 20 may be used to perform the various embodiments described below to implement the encoding method described herein on the encoding side.

[0134] Communication interface 22 can be used to receive encoded image data 21 and transmit the encoded image data 21 via link 13 to destination device 14 or any other device (such as a memory) for storage or direct reconstruction. The other device can be any device used for decoding or storage. Communication interface 22 can, for example, be used to encapsulate the encoded image data 21 into a suitable format, such as data packets, for transmission over link 13.

[0135] Destination device 14 includes decoder 30. Optionally, destination device 14 may also include communication interface 28, image post-processor 32, and display device 34. These are described below:

[0136] Communication interface 28 can be used to receive encoded image data 21 from source device 12 or any other source, such as a storage device, for example, an encoded image data storage device. Communication interface 28 can be used to transmit or receive encoded image data 21 via link 13 between source device 12 and destination device 14 or via any type of network, such as a wired or wireless connection, any type of network, such as a wired or wireless network or any combination thereof, or any type of private and public network, or any combination thereof. Communication interface 28 can be used, for example, to decapsulate data packets transmitted by communication interface 22 to obtain encoded image data 21.

[0137] Both communication interface 28 and communication interface 22 can be configured as unidirectional or bidirectional communication interfaces, and can be used, for example, to send and receive messages to establish connections, acknowledge and exchange any other information related to the communication link and / or data transmission, such as encoded image data transmission.

[0138] Decoder 30 (or video decoder 30) is used to receive encoded image data 21 and provide decoded image data 31 or decoded image 31 (the structural details of decoder 30 will be further described below based on FIG3, FIG4 or FIG5). In some embodiments, decoder 30 can be used to perform the various embodiments described below to implement the decoding method described in this application on the decoding side.

[0139] Image post-processor 32 is used to perform post-processing on decoded image data 31 (also referred to as reconstructed image data) to obtain post-processed image data 33. The post-processing performed by image post-processor 32 may include: color format conversion (e.g., from YUV format to RGB format), color correction, retouching or resampling, or any other processing, and may also be used to transmit the post-processed image data 33 to display device 34.

[0140] Display device 34 is used to receive post-processed image data 33 to display an image to, for example, a user or viewer. Display device 34 can be or may include any class of displays for presenting reconstructed images, such as integrated or external displays or monitors. For example, displays may include liquid crystal displays (LCDs), organic light emitting diode (OLED) displays, plasma displays, projectors, micro-LED displays, liquid crystal on silicon (LCoS), digital light processors (DLP), or any other class of displays.

[0141] Although Figure 1A illustrates source device 12 and destination device 14 as separate devices, device embodiments may also include the functionality of both source device 12 and destination device 14, or both; that is, the functionality of source device 12 or its corresponding features and the functionality of destination device 14 or its corresponding features. In such embodiments, the same hardware and / or software, or separate hardware and / or software, or any combination thereof, may be used to implement the functionality of source device 12 or its corresponding features and the functionality of destination device 14 or its corresponding features.

[0142] Both encoder 20 and decoder 30 can be implemented as any of a variety of suitable circuits, such as one or more microprocessors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), discrete logic, hardware, or any combination thereof. If the technology is implemented in part in software, the device can store the software instructions in a suitable non-transitory computer-readable storage medium, and one or more processors can be used to execute the instructions in hardware to perform the technology of this disclosure. Any of the foregoing (including hardware, software, combinations of hardware and software, etc.) can be considered as one or more processors.

[0143] In some cases, the video encoding and decoding system 10 shown in Figure 1A is merely an example, and the technology of this application can be applied to video encoding setups (e.g., video encoding or video decoding) that do not necessarily involve any data communication between the encoding and decoding devices. In other instances, data may be retrieved from local storage, streamed over a network, etc. The video encoding device may encode the data and store it in storage, and / or the video decoding device may retrieve the data from storage and decode it. In some instances, encoding and decoding are performed by devices that do not communicate with each other but only encode data to storage and / or retrieve data from storage and decode the data.

[0144] Referring to FIG1B, FIG1B is an illustrative diagram of an example of a video decoding system 40 including the encoder 20 of FIG2 and / or the decoder 30 of FIG3 according to an exemplary embodiment. The video decoding system 40 can implement various combinations of technologies of the embodiments of this application. In the illustrated embodiment, the video decoding system 40 may include an imaging device 41, an encoder 20, a decoder 30 (and / or a video encoder / decoder implemented by logic circuitry of a processing unit 46), an antenna 42, one or more processors 43, one or more memories 44, and a display device 45.

[0145] As shown in Figure 1B, the imaging device 41, antenna 42, processing unit 46, logic circuit, encoder 20, decoder 30, processor 43, memory 44, and display device 45 are capable of communicating with each other. As discussed, although encoder 20 and decoder 30 are used as examples to describe the video decoding system 40, in different instances, the video decoding system 40 may contain only encoder 20 or only decoder 30.

[0146] In some instances, antenna 42 can be used to transmit or receive encoded video data streams. Additionally, in some instances, display device 45 can be used to present video data. In some instances, logic circuitry can be implemented using processing unit 46. Processing unit 46 can include an ASIC, graphics processor, general-purpose processor, etc. Video decoding system 40 can also include an optional processor 43, which can similarly include an ASIC, graphics processor, general-purpose processor, etc. In some instances, logic circuitry can be implemented in hardware, such as dedicated video encoding hardware, while processor 43 can be implemented in general-purpose software, operating system, etc. Furthermore, memory 44 can be any type of memory, such as volatile memory (e.g., static random access memory (SRAM), dynamic random access memory (DRAM), etc.) or non-volatile memory (e.g., flash memory, etc.). In a non-limiting instance, memory 44 can be implemented using cache memory. In some instances, logic circuitry can access memory 44 (e.g., for implementing an image buffer). In other instances, the logic circuitry and / or processing unit 46 may include memory (e.g., cache, etc.) for implementing image buffers, etc.

[0147] In some instances, the encoder 20 implemented via logic circuitry may include (e.g., implemented via processing unit 46 or memory 44) an image buffer and (e.g., implemented via processing unit 46) a graphics processing unit. The graphics processing unit may be communicatively coupled to the image buffer. The graphics processing unit may include the encoder 20 implemented via logic circuitry to implement various modules discussed with reference to Figure 2 and / or any other encoder system or subsystem described herein. The logic circuitry may be used to perform various operations discussed herein.

[0148] In some instances, decoder 30 may be implemented via logic circuitry in a similar manner to implement the various modules discussed in reference to decoder 30 of Figure 3 and / or any other decoder system or subsystem described herein. In some instances, the logic circuitry-implemented decoder 30 may include an image buffer (implemented via processing unit 46 or memory 44) and a graphics processing unit (e.g., implemented via processing unit 46). The graphics processing unit may be communicatively coupled to the image buffer. The graphics processing unit may include decoder 30 implemented via logic circuitry to implement the various modules discussed in reference to Figure 3 and / or any other decoder system or subsystem described herein.

[0149] In some instances, antenna 42 can be used to receive an encoded stream of video data. As discussed herein, the encoded stream may contain data related to encoded video frames, indicators, index values, mode selection data, etc., such as data related to code segmentation (e.g., transform coefficients or quantized transform coefficients, optional indicators, and / or data defining code segmentation). Video decoding system 40 may also include a decoder 30 coupled to antenna 42 for decoding the encoded stream. Display device 45 is used to display the video frames.

[0150] It should be understood that, referring to the examples described for encoder 20 in the embodiments of this application, decoder 30 can be used to perform the reverse process. Regarding signaling syntax elements, decoder 30 can be used to receive and parse such syntax elements, and accordingly decode the associated video data. In some examples, encoder 20 can entropy-encode syntax elements into an encoded video stream. In such instances, decoder 30 can parse such syntax elements and accordingly decode the associated video data.

[0151] It should be noted that the encoding and decoding method described in the embodiments of this application is mainly used for the encoding and decoding process of video or images. This process exists in both the encoder 20 and the decoder 30. The encoder 20 and decoder 30 in the embodiments of this application can be, for example, the encoder / decoder corresponding to video standard protocols such as H.263, H.264, HEVV, MPEG-2, MPEG-4, VP8, VP9, ​​or next-generation video standard protocols (such as H.266).

[0152] Referring to Figure 2, which is a schematic / conceptual block diagram of an exemplary example of encoder 20, encoder 20 includes a residual calculation unit 204, a transform processing unit 206, a quantization unit 208, an inverse quantization unit 210, an inverse transform processing unit 212, a reconstruction unit 214, a buffer 216, a loop filter unit 220, a decoded picture buffer (DPB) 230, a prediction processing unit 260, and an entropy coding unit 270. Prediction processing unit 260 may include inter-frame prediction unit 244, intra-frame prediction unit 254, and mode selection unit 262. Inter-frame prediction unit 244 may include a motion estimation unit and a motion compensation unit (not shown). Encoder 20 shown in Figure 2 may also be referred to as a hybrid video encoder or a video encoder based on a hybrid video codec.

[0153] Referring to Figure 3, which is a schematic / conceptual block diagram of an example of a decoder 30, the decoder 30 is used to receive, for example, encoded image data (e.g., encoded bitstream) 21 encoded by encoder 20 to obtain a decoded image 331. During the decoding process, the decoder 30 receives video data from encoder 20, such as encoded video bitstreams representing image blocks of encoded video stripes and associated syntax elements.

[0154] In the example of Figure 3, decoder 30 includes an entropy decoding unit 304, an inverse quantization unit 310, an inverse transform processing unit 312, a reconstruction unit 314 (e.g., a summer 314), a buffer 316, a loop filter 320, a decoded image buffer 330, and a prediction processing unit 360. Prediction processing unit 360 may include an inter-frame prediction unit 344, an intra-frame prediction unit 354, and a mode selection unit 362. In some instances, decoder 30 may perform a decoding process that is generally the inverse of the encoding process described in video encoder 20 of Figure 2.

[0155] Referring to Figure 4, Figure 4 is a structural schematic diagram of a video decoding device 400 (e.g., a video encoding device 400 or a video decoding device 400) provided in an embodiment of this application. The video decoding device 400 is adapted to implement the embodiments described herein. In one embodiment, the video decoding device 400 may be a video decoder (e.g., decoder 30 of Figure 1A) or a video encoder (e.g., encoder 20 of Figure 1A). In another embodiment, the video decoding device 400 may be one or more components of the decoder 30 of Figure 1A or the encoder 20 of Figure 1A.

[0156] The video decoding device 400 includes: an input port 410 and a receiving unit (Rx) 420 for receiving data; a processor, logic unit, or central processing unit (CPU) 430 for processing data; a transmitter unit (Tx) 440 and an output port 450 for transmitting data; and a memory 460 for storing data. The video decoding device 400 may also include photoelectric conversion components and electro-optical (EO) components coupled to the input port 410, receiver unit 420, transmitter unit 440, and output port 450 for the input or output of optical or electrical signals.

[0157] Processor 430 is implemented in both hardware and software. Processor 430 can be implemented as one or more CPU chips, cores (e.g., multi-core processors), FPGAs, ASICs, and DSPs. Processor 430 communicates with ingress port 410, receiver unit 420, transmitter unit 440, egress port 450, and memory 460. Processor 430 includes either an encoding module 470 or a decoding module 470. The encoding / decoding module 470 implements the embodiments disclosed herein to implement the encoding or decoding methods provided in the embodiments of this application. For example, the encoding / decoding module 470 implements, processes, or provides various encoding operations. Therefore, the encoding / decoding module 470 provides a substantial improvement to the functionality of the video decoding device 400 and affects the transitions of the video decoding device 400 to different states. Alternatively, the encoding / decoding module 470 can be implemented with instructions stored in memory 460 and executed by processor 430.

[0158] Memory 460 includes one or more disks, tape drives, and solid-state drives, which can be used as overflow data storage devices to store programs while they are selectively executed, and to store instructions and data read during program execution. Memory 460 can be volatile and / or non-volatile, and can be read-only memory (ROM), random access memory (RAM), ternary content-addressable memory (TCAM), and / or static random access memory (SRAM).

[0159] Referring to FIG5, FIG5 is a simplified block diagram of an apparatus 500 that can be used as either or both of the source device 12 and destination device 14 in FIG1A according to an exemplary embodiment. The apparatus 500 can implement the technology of this application. In other words, FIG5 is a schematic block diagram of an implementation of an encoding or decoding apparatus (hereinafter referred to as decoding apparatus 500) according to an embodiment of this application. The decoding apparatus 500 may include a processor 510, a memory 530, and a bus system 550. The processor and the memory are connected via the bus system. The memory is used to store instructions, and the processor is used to execute the instructions stored in the memory. The memory of the decoding apparatus stores program code, and the processor can call the program code stored in the memory to execute various video encoding or decoding methods described in this application. To avoid repetition, detailed descriptions are not provided here.

[0160] In this embodiment, the processor 510 may be a central processing unit (CPU), or it may be other general-purpose processors, DSPs, ASICs, FPGAs, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0161] The memory 530 may include a read-only memory (ROM) device or a random access memory (RAM) device. Any other suitable type of storage device may also be used as memory 530. Memory 530 may include code and data 531 accessed by processor 510 using bus 550. Memory 530 may further include an operating system 533 and an application program 535, which includes at least one program that allows processor 510 to execute the encoding or decoding methods described in this application. For example, application program 535 may include applications 1 to N, which further include video encoding or decoding applications that execute the encoding or decoding methods described in this application.

[0162] In addition to the data bus, the bus system 550 may also include a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled as bus system 550 in the diagram.

[0163] Optionally, the decoding device 500 may also include one or more output devices, such as a display 570. In one example, the display 570 may be a haptic display that combines a display with a haptic unit capable of operatively sensing touch input. The display 570 may be connected to the processor 510 via a bus 550.

[0164] For example, commonly used transform methods in image coding include discrete cosine transform and wavelet transform. Wavelet transform is a local transform method that can perform localized, multi-scale analysis of images, focusing on the details of signal changes, making it very suitable for image coding tasks.

[0165] Referring to Figure 6, which is an exemplary schematic / conceptual block diagram of a wavelet transform-based encoder, the encoder 60 includes, but is not limited to, a wavelet forward transform unit 610, a quantization unit 620, and an entropy coding unit 630. Specifically, the encoder acquires the original image 601 through an interface unit (also called an input interface, not shown in the figure). The original image 601 is input to the wavelet forward transform unit 610, and after wavelet transform (also called wavelet forward transform), wavelet transform coefficients 602 are obtained, which can also be simply referred to as wavelet coefficients in this embodiment. The quantization unit 620 quantizes the wavelet transform coefficients 602 to obtain quantization coefficients 603. The entropy coding unit 630 entropy codes the quantization coefficients 603 to obtain the image bitstream 604 of the original image, which can also be called a compressed bitstream or bitstream.

[0166] It should be noted that the wavelet transform-based encoder architecture in Figure 6 is for illustrative purposes only, and other instances may include more units or modules. For example, it may include, but is not limited to, prediction units, transform units, etc., and this application does not impose any limitations.

[0167] Referring to Figure 7, which is an exemplary schematic / conceptual block diagram of a wavelet transform-based decoder, the decoder 70 in the example of Figure 7 includes, but is not limited to, an entropy decoding unit 710, an inverse quantization unit 720, and an inverse wavelet transform unit 730. Specifically, at the decoding end, the decoder interfaces with the image bitstream 701 through an input interface (also called an interface unit, not shown in the figure). The entropy decoding unit 710 performs entropy decoding on the image bitstream 701 to obtain quantization coefficients 702. The image bitstream 701 can be a bitstream 604 generated based on the encoder 60 in Figure 6. The inverse quantization unit 720 performs inverse quantization on the quantization coefficients 702 to obtain inverse quantization coefficients 703 (dequantized coefficient(s)), which can also be called reconstructed coefficients, reconstructed wavelet coefficients, etc. The inverse wavelet transform unit 730 performs inverse wavelet transform on the inverse quantization coefficients 703 to obtain the reconstructed image 704.

[0168] It should be noted that the wavelet transform-based decoder architecture in Figure 7 is for illustrative purposes only, and other instances may include more units or modules. For example, it may include, but is not limited to, prediction units, inverse transform units, etc., and this application does not impose any limitations.

[0169] For example, the wavelet coefficients obtained after wavelet transform include wavelet coefficients of the high-frequency subband and wavelet coefficients of the low-frequency subband. In the architecture shown in Figures 6 and 7, the codec performs the same processing on the wavelet coefficients of the high-frequency subband and the low-frequency subband, which has high processing complexity. In the embodiments of this application, the wavelet coefficients of the high-frequency subband can be simply referred to as the high-frequency subband, and the wavelet coefficients of the low-frequency subband can be simply referred to as the low-frequency subband. It can be understood that the high-frequency subband obtained after wavelet transform can optionally be the set of wavelet coefficients of the high-frequency subband, and the low-frequency subband obtained after wavelet transform can optionally be the set of wavelet coefficients of the low-frequency subband.

[0170] For example, the frequency of the high-frequency sub-band is greater than the frequency of the low-frequency sub-band. This can be understood as the frequency of various coefficients corresponding to the high-frequency sub-band in this application (such as wavelet coefficients, reconstruction coefficients, inverse quantization coefficients, etc. involved in this application) being greater than the frequency of the coefficients of the low-frequency sub-band.

[0171] This application provides a wavelet transform-based codec that can independently encode and decode low-frequency and high-frequency sub-bands, effectively reducing encoding and decoding complexity and improving efficiency. For example, an image undergoes wavelet transform to obtain low-frequency and high-frequency sub-bands, which are then encoded to generate low-frequency and high-frequency sub-band bitstreams, respectively. The low-frequency sub-band can be understood as a sub-image representing the low-frequency signal (or low-frequency information) of the original image, and the high-frequency sub-band can be understood as a sub-image representing the high-frequency signal (or high-frequency information) of the original image.

[0172] Referring to Figure 8, which is a schematic / conceptual block diagram of an encoder as an example, the encoder 80 includes, but is not limited to, a sub-graph partitioning unit 810, a wavelet forward transform unit 820, a low-frequency sub-band processing path 830, and a high-frequency sub-band processing path 840. Optionally, the wavelet forward transform unit and the wavelet inverse transform unit in this embodiment can also be collectively referred to as wavelet transform units, which can be understood as performing both wavelet forward transform processing and wavelet inverse transform processing. In this application, the low-frequency sub-band processing path can also be referred to as a low-frequency sub-band processing unit, and the high-frequency sub-band processing path can also be referred to as a high-frequency sub-band processing unit.

[0173] The low-frequency subband processing path 830 includes, but is not limited to: a transform / quantization unit 831 (also known as a low-frequency subband transform / quantization unit) and a low-frequency subband entropy coding unit 832.

[0174] The high-frequency subband processing path 840 includes, but is not limited to: a transform / quantization unit 841 (also known as a high-frequency subband transform / quantization unit) and a high-frequency subband entropy coding unit 842.

[0175] Specifically, encoder 80 receives image 801. Image 801 can be an image in an image sequence that forms a video or video sequence. Image 801 can be referred to as the current image or the image to be encoded, and image 801 can be the original image or an image obtained by processing the original image.

[0176] The sub-image partitioning unit 810 is used to acquire the current image and partition it to obtain at least one sub-image. Specifically, the sub-image partitioning unit 810 partitions the current image into N sub-images according to a sub-image partitioning method, where N is an integer greater than 0 (or an integer greater than 1). The sub-image partitioning method can include, but is not limited to, at least one of the following:

[0177] The width and / or height of the subgraph are multiples of 128;

[0178] The maximum width of the subimage is 1024 pixels;

[0179] The minimum height and / or width of the subimage is 256 pixels;

[0180] The original image resolution is less than or equal to 1080p, and N is an integer greater than 1 and less than or equal to 8; or,

[0181] The original image has a length greater than or equal to 4320 pixels, a width greater than or equal to 2160 pixels, and N is an integer greater than 1 and less than or equal to 16; or,

[0182] The original image has a length greater than or equal to 7680 pixels, a width greater than or equal to 4320 pixels, and N is an integer greater than 1 and less than or equal to 32.

[0183] For example, common video resolutions include: 1280x720, 1280x1080, 1440x1080, 1920x1080 (1080p), 2048x1080, 2048x1556, 3840x2160 (4K), 4096x2160, 5120x2700, 6144x3240, 7680x4320 (8K), and 8192x4320. The above values ​​are only illustrative examples and can be set according to actual needs.

[0184] For example, the subgraph partitioning unit 810 can set specific subgraph specifications (including width and height) based on the above subgraph partitioning method. The specific values ​​can be set according to actual needs within the range specified by the subgraph partitioning method.

[0185] Referring to Figure 9A, which is an exemplary schematic diagram of sub-image partitioning, in the example of Figure 9A, the sub-image partitioning unit 810 can partition the image 801 into m*n sub-images according to the sub-image partitioning method. Optionally, in this example, the width and height of each sub-image satisfy a multiple of 128.

[0186] Referring to Figure 9B, which is an exemplary schematic diagram of sub-image partitioning, in the example of Figure 9B, the sub-image partitioning unit 810 can partition the image into m*n sub-images according to the sub-image partitioning method. When the sub-image partitioning unit 810 partitions the boundary portion of the current image, the width and / or height of the boundary sub-images may optionally not satisfy a multiple of 128. Figure 9B only illustrates this by showing that the width of some boundaries does not satisfy a multiple of 128. Referring to Figure 9B, in this embodiment, for sub-images whose width and height do not satisfy a multiple of 128, the sub-image partitioning unit 810 can pad these sub-images, such as the gray portion in Figure 9B. Taking sub-image 1_n as an example, the height of sub-image 1_n satisfies a multiple of 128, but its width does not. The sub-image partitioning unit 810 pads sub-image 1_n.

[0187] In one possible implementation, the width and / or height of the filled subgraph can optionally be less than or equal to the subgraph size set by the subgraph partitioning method. For example, the size (i.e., width and height) of the filled subgraph 1_n can be the same as that of subgraph 1_1. As another example, the size of the filled subgraph 1_n can be less than the size of subgraph 1_1, where the width of subgraph 1_1 is a, the height is b, the width of subgraph 1_n before filling is c, the height is b, c is less than b, and c is not a multiple of 128. The subgraph partitioning unit 810 fills the subgraph 1_n, and the width of the filled subgraph 1_n is d, and the height is b. Here, d can optionally be a multiple of 16, and d is greater than c and less than or equal to a.

[0188] In this way, by padding the sub-images, each macroblock can meet the 8x8 (pixel) partitioning requirement during the block division process of the encoding. Furthermore, by finding the least common multiple of the width and / or height of the sub-image before padding and 16, sub-images that do not meet the multiple of 16 are padded. This ensures that the size of the sub-image is a multiple of 16 while minimizing the size of the padded sub-image, thereby reducing the encoding complexity of the padded sub-image.

[0189] Optionally, the sub-image partitioning method can be the same for each image in the same video sequence.

[0190] Optionally, the sub-image division unit 810 can divide the image into sub-images along the horizontal and vertical directions, starting from the upper left corner, according to a preset sub-image size. The division order can be set according to actual needs. Typically, the sub-images that need to be filled are boundary sub-images, as shown in Figure 9B.

[0191] In this embodiment, each sub-image is encoded and decoded independently. Specifically, the image is divided into N sub-images, and the encoder 80 encodes each of the N sub-images separately. During the encoding process, a sub-image can also be referred to as the current sub-image or the sub-image to be encoded.

[0192] Referring again to Figure 8, the wavelet forward transform unit 820 is used to perform wavelet transform (also called wavelet forward transform) on the subgraph to obtain low-frequency subband and high-frequency subband. The low-frequency subband includes low-frequency signals in the subgraph that satisfy the low-frequency filter coefficients, and the high-frequency subband includes high-frequency signals in the subgraph that have been decomposed by the high-frequency filter in the wavelet transform.

[0193] Referring to Figure 10, which is an exemplary schematic diagram of wavelet transform, in the example of Figure 10, the wavelet forward transform unit 820 acquires the current sub-image, for example, sub-image 1_1. The wavelet forward transform unit 820 performs a wavelet transform on the current sub-image, wherein the wavelet transform includes one horizontal wavelet transform and one vertical wavelet transform to obtain the wavelet coefficients of the low-low (LL) sub-band (abbreviated as LL sub-band), the wavelet coefficients of the low-high (LH) sub-band (abbreviated as LH sub-band), the wavelet coefficients of the high-high (HH) sub-band (abbreviated as HH sub-band), and the wavelet coefficients of the high-low (HL) sub-band (abbreviated as HL sub-band).

[0194] In this embodiment, the low-frequency subband includes an LL subband, and the high-frequency subband includes an LH subband, an HH subband, and an HL subband. Optionally, the LL subband, LH subband, HH subband, and HL subband have the same dimensions (including width and height).

[0195] Referring again to Figure 8, the low-frequency subband processing path 830 is used to encode the low-frequency subband to obtain low-frequency subband encoded data 806, which can also be referred to as coded low-frequency subband. The high-frequency subband processing path 840 is used to encode the high-frequency subband to obtain high-frequency subband encoded data 809, which can also be referred to as coded high-frequency subband. In this application, the high-frequency subband encoded data can also be referred to as second encoded data.

[0196] Specifically, the wavelet forward transform unit 820 outputs wavelet coefficients 803 of the low-frequency subband to the transform / quantization unit 831, and outputs wavelet coefficients 804 of the high-frequency subband to the transform / quantization unit 841.

[0197] For example, the transform / quantization unit 831 is used to obtain the wavelet coefficients 803 of the low-frequency sub-band, perform quantization processing on the wavelet coefficients 803 of the low-frequency sub-band, or perform transform and quantization processing, and output the quantized coefficients 805 of the low-frequency sub-band, or the quantized wavelet coefficients of the low-frequency sub-band.

[0198] The transform / quantization unit 831 outputs the quantization coefficients 805 of the low-frequency subband to the low-frequency subband entropy encoding unit 832. Optionally, the transform / quantization unit 831 may include, but is not limited to, a transform processing unit and a quantization processing unit (not shown in the figure).

[0199] The transform processing unit is used to transform the low-frequency subband and output transform coefficients, also known as low-frequency subband transform coefficients. The quantization processing unit is used to quantize the subband or the transformed transform coefficients and output quantized coefficients (also known as quantization results).

[0200] The low-frequency subband entropy coding unit 832 is used to acquire the data to be encoded and perform entropy coding on the data to be encoded to obtain low-frequency subband encoded data. The data to be encoded may include, but is not limited to, the quantization coefficients 805 of the low-frequency subband. In some instances, the data to be encoded may also include parameters or information (or syntax elements) input from other modules not shown in the figure. The low-frequency subband encoded data may also be referred to as the first encoded data in this application.

[0201] Referring again to Figure 8, the transform / quantization unit 841 is used to quantize the wavelet coefficients 804 of the high-frequency subband, or to perform both transform and quantization processing, to obtain the quantized coefficients 807 of the high-frequency subband, which can also be referred to as the quantized wavelet coefficients of the high-frequency subband. The transform / quantization unit 841 outputs the quantized coefficients 807 of the high-frequency subband to the high-frequency subband entropy coding unit 842. For parts not described, please refer to the low-frequency subband transform / quantization unit 831; further details are omitted here.

[0202] The high-frequency subband entropy coding unit 842 is used to acquire the data to be encoded and perform entropy coding on the data to be encoded to obtain high-frequency subband encoded data 808, which can also be called encoded high-frequency subband. Optionally, the data to be encoded may include, but is not limited to, the quantization coefficients 807 of the high-frequency subband. In some instances, the data to be encoded may also include parameters or information (or syntax elements) input from other modules not shown in the figure.

[0203] Optionally, the entropy coding unit (including the low-frequency subband entropy coding unit 832 and the high-frequency subband entropy coding unit 842) is used to encode the data to be encoded using an entropy coding algorithm or scheme. The aforementioned entropy coding scheme may be, for example, at least one of the following: variable-length coding (VLC), context-adaptive VLC (CAVLC), arithmetic coding, and context-adaptive binary arithmetic coding (CABAC).

[0204] Optionally, the encoder 80 may also include, but is not limited to, a combining unit (not shown in the figure), which may also be called a multiplexer (MUX). The combining unit is used to generate an image bitstream based on the low-frequency subband coded data 806 and the high-frequency subband coded data 808.

[0205] Specifically, the combining unit writes low-frequency subband coded data 806 into the image bitstream and writes high-frequency subband coded data 808 into the image bitstream. In this embodiment, by encoding the low-frequency subband and high-frequency subband separately, the low-frequency subband coded data and high-frequency subband coded data can be decoded independently. That is, at the decoding end, it can independently decode the low-frequency subband coded data and high-frequency subband coded data in the image bitstream, thereby improving decoding efficiency.

[0206] As mentioned above, a single image can be divided into multiple sub-images, for example, N sub-images. Each sub-image can be decomposed into wavelet coefficients of a high-frequency sub-band and wavelet coefficients of a low-frequency sub-band after wavelet forward transform. Encoder 80 independently encodes the low-frequency and high-frequency sub-bands of each sub-image to obtain the low-frequency sub-band encoded data and high-frequency sub-band encoded data of a single sub-image. In this way, the combining unit can obtain the high-frequency sub-band encoded data and low-frequency sub-band encoded data of each of the N sub-images, that is, obtain N high-frequency sub-band encoded data (optionally, the high-frequency sub-band encoded data of each sub-image includes the HL sub-band encoded data corresponding to the HL sub-band, the HH sub-band encoded data corresponding to the HH sub-band, and the LH sub-band encoded data corresponding to the LH sub-band) and N low-frequency sub-band encoded data. The combining unit can optionally write the low-frequency sub-band encoded data and high-frequency sub-band encoded data of the N sub-images into the image bitstream. The format (or structure) of the image bitstream will be described in detail below.

[0207] Referring to Figure 11, which is a schematic / conceptual block diagram of an exemplary decoder, decoder 110 is used to receive, for example, a bitstream encoded by an encoder to obtain a decoded image, also referred to as a reconstructed image 1108. During the decoding process, the decoder receives a bitstream (also referred to as the bitstream of the original image, or the image data of the original image) from the encoder.

[0208] In the example shown in Figure 11, the decoder 110 includes, but is not limited to: a high-frequency subband processing path 1120, a low-frequency subband processing path 1110, a wavelet inverse transform unit 1130, and a subgraph stitching unit 1140.

[0209] The low-frequency subband processing path 1110 includes, but is not limited to: a low-frequency subband entropy decoding unit 1111 and an inverse quantization / inverse transform unit 1112 (also referred to as a low-frequency subband inverse quantization / inverse transform unit). The high-frequency subband decoding path 1120 includes, but is not limited to: a high-frequency subband entropy decoding unit 1121 and an inverse quantization / inverse transform unit 1122 (also referred to as a high-frequency subband inverse quantization / inverse transform unit).

[0210] For example, decoder 110 receives an image bitstream and obtains a reconstructed image 1108 (also known as a decoded image) based on high-frequency subband coded data 1102 and low-frequency subband coded data 1101 in the image bitstream.

[0211] For example, the low-frequency subband entropy decoding unit 1111 is used to acquire low-frequency subband encoded data 1101 in the image bitstream to obtain entropy decoding data, which includes, but is not limited to, the quantization coefficients 1103 of the low-frequency subband. Specifically, the low-frequency subband entropy decoding unit 1111 performs entropy decoding processing on the low-frequency subband encoded data 1101 to obtain entropy decoding data. The low-frequency subband entropy decoding unit 1111 outputs the quantization coefficients 1103 of the low-frequency subband to the inverse quantization / inverse transform unit 1112.

[0212] The inverse quantization / inverse transform unit 1112 is used to obtain the quantization coefficients 1103 of the low-frequency subband to obtain the reconstruction coefficients 1105 of the low-frequency subband, which can also be simply referred to as the reconstructed low-frequency subband. In some instances, it can also be called the reconstructed value of the low-frequency subband, the reconstruction coefficient of the low-frequency subband, or the reconstructed wavelet coefficients of the low-frequency subband. Specifically, the low-frequency inverse quantization / inverse transform unit performs inverse quantization and / or inverse transform processing on the quantization coefficients to obtain the reconstructed low-frequency subband, such as the reconstructed LL subband. Optionally, the inverse quantization / inverse transform unit 1112 may include, but is not limited to, an inverse quantization processing unit and an inverse transform processing unit (not shown in the figure). The inverse quantization processing unit can be used to perform inverse quantization processing on the input parameters to obtain inverse quantization coefficients. The inverse transform unit can be used to perform inverse transform processing on the input parameters to obtain inverse transform coefficients, which can also be called inverse transform coefficients. For example, the inverse transform processing unit can perform inverse transform processing on the inverse quantization coefficients output by the inverse quantization processing unit to obtain inverse transform inverse quantization coefficients.

[0213] The high-frequency subband entropy decoding unit 1121 is used to acquire the high-frequency subband encoded data 1102 in the image bitstream to obtain entropy-decoded data. The entropy-decoded data includes, but is not limited to, the quantization coefficients 1104 of the high-frequency subband. Specifically, the high-frequency subband entropy decoding unit 1121 performs entropy decoding processing on the high-frequency subband encoded data 1102 to obtain the entropy-decoded data. The high-frequency subband entropy decoding unit 1121 outputs the quantization coefficients 1104 of the high-frequency subband to the inverse quantization / inverse transform unit 1122.

[0214] The inverse quantization / inverse transform unit 1122 is used to obtain the quantization coefficients 1104 of the high-frequency subband, resulting in the reconstructed coefficients 1106 of the high-frequency subband, which can be simply referred to as the reconstructed high-frequency subband. In some instances, it can also be called the reconstructed value of the high-frequency subband, the reconstructed wavelet coefficients of the high-frequency subband, etc. The inverse quantization / inverse transform unit 1122 outputs the reconstructed high-frequency subband to the wavelet inverse transform unit 1130. Specifically, the inverse quantization / inverse transform unit 1122 performs inverse quantization processing on the quantization coefficients 1104 of the high-frequency subband, or performs inverse quantization and inverse transform processing, to obtain the reconstructed high-frequency subband. For example, this includes reconstructing the HH subband (e.g., the reconstructed coefficients of the HH subband), reconstructing the HL subband (e.g., the reconstructed coefficients of the HL subband), and reconstructing the LH subband (e.g., the reconstructed coefficients of the LH subband).

[0215] For example, the wavelet inverse transform unit 1130 is used to obtain the reconstruction coefficients 1105 of the low-frequency sub-band and the reconstruction coefficients 1106 of the high-frequency sub-band to obtain the reconstructed sub-map 1107. Specifically, the wavelet inverse transform unit 1130 can obtain the reconstructed low-frequency sub-band and the reconstructed high-frequency sub-band of each sub-map. The wavelet inverse transform unit performs wavelet inverse transform processing on the obtained reconstructed low-frequency sub-band and reconstructed high-frequency sub-band corresponding to the same sub-map to obtain the reconstructed sub-map. Its inverse transform process is the inverse of the wavelet forward transform process, which will not be described in detail here.

[0216] The sub-image stitching unit 1140, also known as an image reconstruction unit or sub-image combination unit, is used to obtain reconstructed sub-images of the original image to obtain a reconstructed image of the original image. Specifically, the sub-image stitching unit 1140 obtains N reconstructed sub-images corresponding to the current decoded image. Based on the N reconstructed sub-images, the sub-image stitching unit 1140 can obtain the reconstructed image (also known as the decoded image or the decoded image) of the current image.

[0217] Based on the codecs shown in Figures 8 and 11, there can be more variations of the codec, such as including more processing units or modules.

[0218] Referring to Figure 12, which is a schematic / conceptual block diagram of an encoder as an example, the encoder in the example of Figure 12 includes, but is not limited to, a sub-graph partitioning unit 1210, a wavelet forward transform unit 1220, a low-frequency sub-band processing path 1230, and a high-frequency sub-band processing path 1240.

[0219] The descriptions of the subgraph partitioning unit 1210 and the wavelet forward transform unit 1220 can be found in the relevant content in Figure 8, and will not be repeated here.

[0220] The low-frequency subband processing path 1230 is used to obtain the wavelet coefficients 1203 of the low-frequency subband to obtain the low-frequency subband encoded data 1213. The low-frequency subband processing path 1230 includes, but is not limited to: a block partitioning unit 1231 (also called the low-frequency subband block partitioning unit 1231 or the first block partitioning unit 1231), a residual calculation unit 1232, a prediction unit 1237, a control unit 1238, a transform / quantization unit (also called the low-frequency subband transform / quantization unit 1233 or the first transform / quantization unit), an inverse quantization / inverse transform unit 1234 (also called the low-frequency subband inverse quantization / inverse transform unit 1234 or the first inverse transform / inverse quantization unit), a low-frequency subband wavelet coefficient 1203 splicing unit 1246, a low-frequency subband wavelet coefficient 1203 splicing unit 1246, and a low-frequency subband entropy coding unit 1239, etc.

[0221] The high-frequency subband processing path 1240 is used to acquire high-frequency subbands to obtain high-frequency subband encoded data. The high-frequency subband processing path 1240 includes, but is not limited to: block partitioning unit 1231 (also referred to as high-frequency subband block partitioning unit 1231 or second block partitioning unit 1231), transform / quantization unit (also referred to as transform / quantization unit 1242 or second transform / quantization unit), high-frequency subband entropy coding unit 1243, etc.

[0222] Alternatively, in some instances, the encoder may include more or fewer units or modules than in the structure shown in Figure 12.

[0223] The image 1201 encoding method provided in this application will be described in detail below with reference to the encoder shown in Figure 12:

[0224] The codec receives image 1201. A description of image 1201 can be found above and will not be repeated here.

[0225] Sub-image partitioning unit 1210 partitions image 1201 into sub-images and outputs N sub-images. N is an integer greater than 0. In this embodiment, each sub-image is encoded and decoded independently. During the encoding process, sub-image 1202 can be referred to as the current sub-image or the sub-image to be encoded.

[0226] Wavelet forward transform unit 1220 performs wavelet forward transform on the current sub-image to obtain wavelet coefficients 1203 (hereinafter referred to as low-frequency sub-band) and wavelet coefficients 1214 (hereinafter referred to as high-frequency sub-band) of the current sub-image. The wavelet coefficients 1203 of the low-frequency sub-band include the wavelet coefficients of the LL sub-band, and the high-frequency sub-band includes the wavelet coefficients of the LH, HL, and HH sub-bands. In this embodiment, each sub-image of the image can be independently encoded and decoded, and the high-frequency sub-band and low-frequency sub-band of each sub-image are independently encoded and decoded. The LH, HL, and HH sub-bands in the high-frequency sub-band can also be independently encoded and decoded.

[0227] The block partitioning unit 1231 (which may be called the low-frequency sub-band block partitioning unit) is used to obtain the wavelet coefficients 1203 of the low-frequency sub-band of the current sub-graph 1202, so as to obtain at least one macroblock 1204 of the low-frequency sub-band of the sub-graph 1202, wherein the macroblock can also be understood as a set of partial coefficients in the wavelet coefficients of the low-frequency sub-band.

[0228] Specifically, the block partitioning unit 1231 partitions the wavelet coefficients 1203 of the low-frequency sub-band of the current subgraph 1202 into blocks based on the block partitioning method, obtaining at least one macroblock 1204 of the low-frequency sub-band of the current subgraph, for example, M macroblocks, where M is an integer greater than 0 (or greater than 1). The low-frequency block partitioning unit 1231 outputs the macroblocks 1204 of the wavelet coefficients 1203 of the current low-frequency sub-band one by one to the residual calculation unit 1232 and the control unit 1238.

[0229] In this embodiment, macroblock 1204 is a basic encoding / decoding unit. During the encoding process, macroblock 1204 can also be referred to as the current block, current image 1201 block, macroblock 1204 to be encoded, block to be encoded, image 1204 block to be encoded, etc.

[0230] Alternatively, the block partitioning method includes, but is not limited to:

[0231] The wavelet coefficients 1203 of both the high-frequency subband and the low-frequency subband are divided into basic coding units of 8x8 macroblocks 1204 (unit is pixels).

[0232] For example, as described above, each subband uses macroblock 1204 as the basic coding unit. The macroblock 1204 currently to be encoded is referred to as the current macroblock 1204. Specifically, the low-frequency subband processing path 1230 encodes each macroblock 1204 of the wavelet coefficients 1203 of the low-frequency subband block by block. For example, encoding and prediction are performed on each macroblock 1204. The following description only uses the encoding process of the current macroblock 1204; the processing processes for other macroblocks are the same, and will not be described in detail here. For example, in encoding, it refers to the macroblock currently being encoded; in decoding, it refers to the macroblock currently being decoded. The decoded macroblock in the reference image used for predicting the current macroblock 1204 is called the reference block (i.e., the low-frequency subband reconstruction block 1209 in the figure). The reference block is the block that provides the reference signal for the current block, where the reference signal represents the pixel value within the macroblock 1204. The block in the reference image that provides the prediction signal for the current block can be called prediction block 1205, where the prediction signal represents the pixel value, sample value, or sample signal within prediction block 1205. For example, after traversing multiple reference blocks, an optimal reference block is found, which will provide the prediction for the current block; this block is called prediction block 1205.

[0233] Specifically, referring to Figure 12, the residual calculation unit 1232 is used to obtain the current macroblock 1204 and the prediction block 1205 (further details of the prediction block 1205 are provided below) to obtain the residual block 1206. Specifically, the residual calculation unit performs residual calculations on the current macroblock 1204 and the prediction block 1205 to obtain the residual block 1206. The residual calculation unit 1232 outputs the residual block 1206 to the transform / quantization unit 1233.

[0234] The transform / quantization unit 1233 is used to acquire the residual block 1206 to obtain the residual coefficients 1207. Specifically, the transform / quantization unit 1233 performs transform and / or quantization processing on the residual block 1206 to obtain the residual coefficients 1207, which can also be called the quantization coefficients of the residual block, or the quantized residual block. The transform / quantization unit 1233 outputs the residual coefficients 1207 to the inverse quantization single / inverse transform unit 1234 and the low-frequency subband entropy coding unit 1239.

[0235] The inverse quantization / inverse transform unit 1234, also known as the inverse quantization / inverse transform unit, is used to obtain the residual coefficients 1207 to obtain the residual reconstruction block 1208. Specifically, the inverse quantization / inverse transform unit 1234 performs inverse quantization and / or inverse transform processing on the residual coefficients 1207 to obtain the residual reconstruction block 1208, which can also be referred to as the inverse quantization coefficients of the residual block, the inverse quantized residual block, etc. The inverse quantization / inverse transform unit 1234 outputs the residual reconstruction block 1208 to the low-frequency subband splicing unit 1236.

[0236] The dequantization / inverse transform unit 1234 may include a dequantization unit and an inverse transform unit (not shown in the figure). The dequantization unit is used to dequantize the input coefficients, and the inverse transform unit is used to inverse transform the input coefficients.

[0237] The low-frequency subband reconstruction unit 1235 is used to obtain a low-frequency subband reconstruction block 1209 based on the prediction block 1205 and the residual reconstruction block 1208. Specifically, the low-frequency subband reconstruction unit 1235 adds the residual reconstruction block 1208 to the prediction block 1205 to obtain the low-frequency subband reconstruction block 1209, which can also be referred to as the reconstructed low-frequency subband macroblock. Optionally, the low-frequency subband reconstruction unit 1235 outputs the low-frequency subband reconstruction block 1209 to the prediction unit 1237 and the low-frequency subband splicing unit 1236. Optionally, the low-frequency subband reconstruction unit 1235 outputs the low-frequency subband reconstruction block 1209 to the control unit 1238.

[0238] The low-frequency subband stitching unit 1236 is used to obtain the reconstructed low-frequency subband 1211, which can also be referred to as the reconstructed value or reconstructed data of the low-frequency subband, based on the low-frequency subband reconstruction block 1209. Optionally, the low-frequency subband stitching unit 1236 outputs the reconstructed low-frequency subband 1211 to the prediction unit 1237. Optionally, the low-frequency subband stitching unit 1236 outputs the reconstructed low-frequency subband 1211 to the control unit 1238.

[0239] Specifically, as described above, the low-frequency subband uses macroblocks as the basic coding unit, and the low-frequency subband splicing unit 1236 can obtain M low-frequency subband reconstruction blocks of a low-frequency subband. The low-frequency subband splicing unit 1236 can reconstruct the corresponding low-frequency subband based on the M low-frequency subband reconstruction blocks, that is, obtain the reconstructed low-frequency subband 1211.

[0240] The control unit 1238, also known as the mode selection unit, is used to determine the syntax element 1213 based on the macroblock 1204 (i.e. the current block); or to determine the syntax element 1213 based on the current macroblock 1204, the low-frequency subband reconstruction block 1209, and the reconstructed low-frequency subband 1211.

[0241] Syntax element 1213 includes at least one syntax element, such as pattern information. Pattern information, also known as prediction mode information, is used to indicate the prediction mode (or prediction method) of prediction unit 1237, such as inter-frame or intra-frame prediction mode. Control unit 1238 can output syntax element 1213 to prediction unit 1237 and low-frequency subband entropy coding unit 1239.

[0242] Prediction unit 1237, also known as prediction processing unit, is used to acquire syntax element 1213 and perform prediction processing based on syntax element 1213. Specifically, prediction unit 1237 may select a prediction mode based on syntax element 1213 (e.g., mode information in the syntax element). In one example, prediction unit 1237 may acquire low-frequency subband reconstruction block 1209 based on syntax element 1213 to acquire prediction block 1205. Specifically, prediction unit 1237 may perform intra-frame prediction based on low-frequency subband reconstruction block 1209 to acquire prediction block 1205. In another example, prediction unit 1237 may acquire reconstructed low-frequency subband 1211 based on syntax element 1213 to acquire prediction block 1205.

[0243] The prediction unit 1237 outputs prediction block 1205 to the residual calculation unit 1232 and the low-frequency sub-band splicing unit 1236.

[0244] The low-frequency subband entropy coding unit 1239 is used to obtain low-frequency subband encoded data 1213, also known as encoded low-frequency subband, based on residual coefficients 1207 and syntax elements 1213. Specifically, the low-frequency subband entropy coding unit 1239 uses an entropy coding algorithm or scheme (e.g., variable length coding (VLC) scheme, context adaptive VLC (CAVLC) scheme, arithmetic coding scheme, context adaptive binary arithmetic coding (CABAC), syntax-based context-adaptive binary arithmetic coding (SBAC), probability interval partitioning entropy (PIPE) coding, or other entropy coding methods or techniques) to entropy encode the residual coefficients 1207 and syntax elements 1213 to obtain low-frequency subband encoded data 1213 output in the form of, for example, an encoded bitstream.

[0245] Referring again to Figure 12, the block partitioning unit 1241, also known as the high-frequency subband block partitioning unit, is used to obtain the high-frequency subband of the current subgraph 1202 to obtain at least one macroblock 1215 of the high-frequency subband of the subgraph 1202. For a detailed description, please refer to the low-frequency subband section; it will not be repeated here. Specifically, the block partitioning unit 1241 partitions the high-frequency subband 1214 of the current subgraph 1202 (hereinafter referred to as the current high-frequency subband) into blocks based on the block partitioning method, obtaining at least one macroblock 1215 of the current subgraph 1202, for example, M macroblocks, where M is an integer greater than 0 (or an integer greater than 1). Other undescribed parts can be referred to the relevant description of the block partitioning unit 1231; it will not be repeated here.

[0246] Block partitioning unit 1241 outputs the macroblocks of the current high-frequency subband one by one to quantization / conversion unit 1242.

[0247] The transform / quantization unit 1242 is used to transform and / or quantize the macroblock 1215 to obtain the quantization coefficients 1216 of the high-frequency subband block (i.e., the quantization coefficients of the current macroblock). The transform / quantization unit 1242 outputs the quantization coefficients 1216 of the high-frequency subband block to the high-frequency subband entropy coding unit 1243.

[0248] The high-frequency subband entropy coding unit 1243 is used to perform entropy coding on the data to be encoded to obtain high-frequency subband encoded data 1217. The data to be encoded may include, but is not limited to, the quantization coefficients and syntax elements of each high-frequency subband block. The high-frequency subband encoded data 1217 includes, but is not limited to, HH subband encoded data, HL subband encoded data, and LH subband encoded data.

[0249] The encoding and decoding method provided in this application supports two scenarios: full I-frame configuration and I / P frame alternating encoding configuration. The encoder architecture shown in Figure 12 adds relevant modules required for the prediction process based on the wavelet transform architecture shown in Figure 8, which can improve the compression efficiency of I / P frame alternating encoding for scenarios such as fixed camera positions and slow camera movement.

[0250] Optionally, the encoder may also include a combination unit (not shown in the figure) for generating an image bitstream based on low-frequency subband coded data and high-frequency subband coded data. The specific process can be referred to the relevant description in Figure 8 above.

[0251] The bitstream output by the encoder in the embodiments of this application will be described in detail below. The bitstream structure described below can be applied to the encoders shown in Figures 8 and 12, and of course, it can also be applied to other encoder variations based on Figures 8 or 12.

[0252] For example, as described above, the low-frequency subband entropy coding unit 1239 and the high-frequency subband entropy coding unit 1243 output low-frequency subband coded data 1213 and high-frequency subband coded data 1217, respectively. This can be understood as the encoder 120 independently encoding the low-frequency and high-frequency subbands of each subgraph, outputting low-frequency and high-frequency subband coded data corresponding to the current subgraph.

[0253] Referring to Figure 13A, which is an exemplary schematic diagram of an image bitstream structure, the image bitstream in the example of Figure 13A includes, but is not limited to, image header information and image data (also referred to as an image data region).

[0254] For example, the image data includes at least one image data region (also referred to as an image data sub-region), such as, but not limited to, a first image data region and a second image data region. During the encoding process, the encoder (e.g., through a combination unit) writes high-frequency subband encoded data and low-frequency subband encoded data into the image bitstream. Specifically, the encoder writes high-frequency subband encoded data into the first image data region and low-frequency subband encoded data into the second image data region.

[0255] For example, image header information includes, but is not limited to, offset information and image size information.

[0256] For example, image size information is used to indicate the size of the original image. As mentioned above, during the encoding process, some sub-images may be padded during sub-image partitioning to ensure that the length and width of each sub-image are multiples of 16. Thus, during decoding, the size of the reconstructed image obtained by the decoder may be larger than the original image size. The decoder can process the reconstructed image based on the image size information to remove the padded portions.

[0257] For example, offset information is used to indicate the position of a data region in the image bitstream, and can also be understood as indicating the position of independently decodeable coded data in the image bitstream. When decoding coded data (i.e., the image bitstream) according to this application, the offset information in the image header information can be used to obtain independently decodeable coded data, and decoding operations can be performed on the coded data. The independently decodeable coded data (e.g., low-frequency subband coded data and high-frequency subband coded data) can be decoded synchronously during decoding to improve decoding efficiency.

[0258] In one example, the offset information can be the length of the image data region containing adjacent, independently decodeable encoded data in the image bitstream.

[0259] In another example, the offset information can be the offset (i.e., the difference) between the starting position of the image data region where the independently decoded encoded data is located and the ending position of the image header information.

[0260] Referring to Figure 13B, which is an exemplary schematic diagram of an image bitstream structure, the bitstream in the example of Figure 13B includes, but is not limited to, image header information and image data. The image header information includes, but is not limited to, offset information, etc., detailed concepts of which can be found above and will not be repeated here. The image data includes, but is not limited to, low-frequency subband coded data and high-frequency subband coded data. The low-frequency subband coded data is written to the first image data area, and the high-frequency subband coded data is written to the second image data area. Other descriptions can be found in Figure 13A and will not be repeated here.

[0261] It should be noted that the embodiments in this application only use the image bitstream of a single image as an example for illustration, that is, the bitstream includes only one image data. In the process of encoding video images, the encoder can generate an encoded image bitstream for each image, that is, the image bitstream includes multiple image data, and each image data carries the encoded data of the corresponding image.

[0262] Referring to Figure 14A, which is an exemplary schematic diagram of an image bitstream structure, in the example of Figure 14A, as described above, the high-frequency subband coding data of each sub-image further includes: HH subband coding data, HL subband coding data, and LH subband coding data. Accordingly, in this example, the image data includes, but is not limited to: HH subband coding data, HL subband coding data, and LH subband coding data of each sub-image of the image.

[0263] For example, in the example shown in Figure 14A, LL subband coded data is written to the second image data region, HL subband coded data is written to the first image data subregion, HH subband coded data is written to the second image data subregion, and LH subband coded data is written to the dotted image data subregion. The writing order of each high-frequency subband is only an illustrative example and will not be repeated below.

[0264] Optionally, in the example shown in Figure 14A, the offset information of each segment of independently decodeable encoded data in the image header information may include, but is not limited to: the offset of the starting position of the first image data sub-region relative to the ending position of the image header information (this data is usually measured in n bytes), the offset of the starting position of the second image data sub-region relative to the ending position of the image header information, and the offset of the starting position of the third image data sub-region relative to the ending position of the image header information. After obtaining these offsets by decoding the image header information, the starting position of each segment of independently decodeable encoded data can be obtained, so that the LL subband encoded data, HH subband encoded data, HL subband encoded data, and LH subband encoded data can be decoded independently during decoding.

[0265] Optionally, in the example shown in Figure 14A, the offset information may include, but is not limited to: the length information L0 of the first image data region (this length is typically measured in n bytes), the length information L1 of the first image data sub-region, and the length information L2 of the second image data sub-region. L0 is the offset of the starting position of the first image data sub-region relative to the ending position of the image header information. L0 + L1 yields the offset of the starting position of the second image data sub-region relative to the ending position of the image header information. L0 + L1 + L2 yields the offset of the starting position of the third image data sub-region relative to the ending position of the image header information. After obtaining the offset information by decoding the image header information, the starting position of each independently decodeable segment of encoded data can be obtained by addition calculation, so that the LL subband encoded data, HH subband encoded data, HL subband encoded data, and LH subband encoded data can be decoded independently during decoding.

[0266] Referring to Figure 14B, which is an exemplary schematic diagram of an image bitstream structure, in the example of Figure 14B, low-frequency subband coded data is written to the first image data region, and high-frequency subband coded data is written to the second image data region. Further details can be found in Figure 14A, and will not be repeated here.

[0267] Referring to Figure 15, which is an exemplary schematic diagram of an image bitstream structure, in the example of Figure 15, the high-frequency subband encoded data is written to the corresponding image data region according to the type of subband. Specifically, the description of the first image data region and its sub-regions, and the second image data region, can be referred to above and will not be repeated here. Taking HL subband encoded data as an example, the HL subbands of each sub-image (e.g., sub-image 1 to sub-image N) of the image are written to the first image data sub-region; the order is only illustrative.

[0268] In this context, for example, sub-image 1-HL in the figure represents the HL subband encoded data of sub-image 1 of the image. The HL subband encoded data of each sub-image further includes the encoded macroblocks of the HL subband of that sub-image. For example, sub-image 1-HL includes, but is not limited to: sub-image 1-HL-MB0 to sub-image 1-HL-MBm. Sub-image 1-HL-MBx represents the encoded macroblock MBx in the HL subband of sub-image 1. The HH subband encoded data and LH subband encoded data are similar to the HL subband encoded data and will not be described further here.

[0269] Referring again to Figure 15, the LL subband coding data includes, but is not limited to, the LL subband coding data of each sub-image of the image. For example, sub-image 1-LL to sub-image N-LL. Sub-image 1-LL represents the LL subband coding data of sub-image 1 of the image. Each sub-image-LL further includes, but is not limited to, the coding data of each macroblock of that sub-image, i.e., the coded macroblock. For example, sub-image 1-LL includes, but is not limited to, sub-image 1-LL-MB0 to sub-image 1-LL-MBn. Sub-image 1-LL-MBx represents the coding data of macroblock MBx of the LL subband of sub-image 1 of the image.

[0270] For example, during the encoding process, each sub-band of each subgraph is encoded independently; correspondingly, during the decoding process, each encoded sub-band of each subgraph can be decoded independently. In the example shown in Figure 15, it can be understood that the four encoded sub-bands of each subgraph can be decoded independently. For instance, during decoding, the decoder can obtain the HL sub-band encoded data, LH sub-band encoded data, HH sub-band encoded data, and LL sub-band encoded data of subgraph 1 based on the offset information, and perform independent decoding to obtain the decoded subgraph 1.

[0271] In this example, the coded subbands and their coded macroblocks of each sub-image in the figure can also be written into the corresponding image data sub-regions, which are not shown in the figure and will not be repeated below. Accordingly, offset information can be used to indicate the position of the image region to which the independently coded subbands of each sub-image belong. Thus, during decoding, the decoder can obtain the positions of the four coded subbands (LL subband coded data, HH subband coded data, HL subband coded data, and LH subband coded data) of a single sub-image in the image data based on the offset information, and perform decoding on them to obtain the decoded subbands of the corresponding sub-image (including LL subband decoded data, HH subband decoded data, HL subband decoded data, and LH subband decoded data). For example, during decoding, the decoder obtains sub-image 1-HL (including each coded macroblock contained in its sub-region, the same below, and will not be repeated), sub-image 1-HH, sub-image 1-LH, and sub-image 1-LL. In this way, the decoder can obtain the decoded sub-graph 1 based on the above encoded sub-bands without having to decode other sub-graphs one by one according to the order of the bitstream. This can improve decoding efficiency while reducing the occupation of the decoding buffer (e.g., DPB) and reducing the hardware processing pressure and storage burden on the decoding end.

[0272] In one possible implementation, on the encoding side, the low-frequency subband entropy encoding unit can output low-frequency subband encoded data (e.g., LL subband bitstream), and the high-frequency subband entropy encoding unit can output high-frequency subband encoded data (e.g., including HH subband decoded data, HL subband decoded data, and LH subband decoded data). The combining unit writes this data into the image data to generate an image bitstream.

[0273] In another possible implementation, the low-frequency subband entropy coding unit and the high-frequency subband entropy coding unit can also output the coded macroblock of the coded subband to the combining unit after each macroblock is encoded. The combining unit can generate the bitstream structure as shown in the figures (e.g., Figures 15, 16A-16C, 17A-17B, etc.) according to the specified order of the coded macroblocks of the image data.

[0274] Optionally, the subgraph and bitstream structure in Figure 15 can also be applied to the bitstream structures shown in Figures 13B and 14B, and will not be illustrated in detail here.

[0275] In the embodiments of this application, the subbands in the high-frequency subband coded data of the image data can be interleaved and sorted. The interleaving and sorting can be at the sub-image granularity, the subband type granularity, or the macroblock (MB) granularity. Several interleaving and sorting methods are provided below. It should be noted that the interleaving methods shown in the embodiments of this application are only illustrative examples. In other embodiments, other interleaving methods can also be set according to the encoding and decoding requirements.

[0276] Referring to Figure 16A, which is an exemplary schematic diagram of an image bitstream structure, the interleaving method in the example of Figure 16A is interleaving at the subband type of the subimage. Specifically, as shown in Figure 16A, the HL subband encoded data, encoded HH, and encoded LH of each subimage of the image are continuously written into the image bitstream. The subimage order and macroblock order shown in the figure are merely illustrative examples. The format of the low-frequency subband is not shown in the figure; please refer to Figure 15 and its related description, which will not be repeated here.

[0277] For example, as shown in Figure 16A, sub-images 1-HL, 1-HH, and 1-LH are continuously written into the image bitstream. Sub-image 1-HL represents the HL subband encoded data of sub-image 1, which includes, but is not limited to, all encoded macroblocks (MB) of the HL subband of sub-image 1, such as MB0-HL to MBm-HL. Only the bitstream format of sub-image 1 is shown in the figure; other sub-images are similar and will not be described individually here. The LL subband encoded data can be referred to above and will not be repeated here.

[0278] In this example, similar to the description in Figure 15, each coded subband of each sub-image can be decoded independently. Accordingly, offset information can be used to indicate the location of the image region to which the independently coded subband of each sub-image belongs. This allows, during decoding, the positions of the four coded subbands (LL subband coded data, HH subband coded data, HL subband coded data, and LH subband coded data) of a single sub-image can be obtained in the image data based on the offset information, and decoding can be performed on them to obtain the decoded subbands of the corresponding sub-image (including LL subband decoded data, HH subband decoded data, HL subband decoded data, and LH subband decoded data). Details not described herein can be found in Figure 15 and will not be repeated here.

[0279] Referring to Figure 16B, which is an exemplary schematic diagram of the image bitstream structure, the interleaving method in the example of Figure 16B is interleaving at the macroblock level for each sub-image.

[0280] Specifically, in the example shown in Figure 16B, the high-frequency subband encoded data (including LH subband encoded data, HH subband encoded data, and HL subband encoded data) of each sub-image of the image are continuously written into the first image data region. A description of the second image data region can be found in Figure 15, and will not be repeated here.

[0281] For example, as shown in Figure 16B, sub-images 1-HL-MB0, 1-HH-MB0, and 1-LH-MB0 are consecutively written into the first image data region. Here, 1-HL-MB0 represents the encoded macroblock MB0 of the HL subband of 1, 1-HH-MB0 represents the encoded macroblock MB0 of the HH subband of 1, and 1-LH-MB0 represents the encoded macroblock MB0 of the LH subband of 1. The figure only shows the encoded data structure of 1 in the bitstream; the other sub-images are similar and will not be illustrated individually here.

[0282] In this example, during decoding, the decoding end can decode the high-frequency subband encoded data according to the sub-image order, that is, each sub-image in the first image data region is decoded independently. The low-frequency subband encoded data is also decoded according to the sub-image order, that is, each sub-image in the second image data region is decoded independently. Unlike Figures 15, 16A, and 16C, when the bitstream structure in the above figures is applied to decoding, the decoding end decodes the different types of encoded subbands of each sub-image separately. This can also be understood as needing to simultaneously decode the four sub-bitstreams (each sub-bitstream corresponds to one type of encoded subband) of each sub-image to obtain four decoded subbands of an image at the output end, thereby obtaining the decoded sub-image and storing additional data. For the bitstream structure shown in Figure 16B, the decoding end only needs to decode the high-frequency subband and low-frequency subband of each sub-image, that is, two bitstreams. For example, as shown in Figure 16B, when the decoding end decodes the first image data region, it can decode each encoded macroblock one by one according to the encoded macroblock order of each sub-image in the region. That is, the three high-frequency subband encoded data of sub-Figure 1 are written continuously into the first image data area. Therefore, during decoding, the three high-frequency subband encoded data of sub-Figure 1 can be decoded one by one to obtain the decoded high-frequency subband. The processing of LL subband encoded data can be referred to Figure 15, and will not be described in detail here.

[0283] In the example shown in Figure 16B, the offset information is used to indicate the position of the first image data region and the position of the second image data region. That is, the example shown in Figure 16B includes two sub-bitstreams that can be decoded independently. Compared with the framework and decoding methods in Figures 15, 16A, and 16C, it has lower hardware performance requirements, only requiring simultaneous decoding of the two sub-bitstreams of the sub-image (corresponding to low-frequency sub-band encoded data and high-frequency sub-band encoded data).

[0284] Referring to Figure 16C, which is an exemplary schematic diagram of an image bitstream structure, the interleaving method in Figure 16C is based on the sub-bands of each sub-image. Specifically, as shown in Figure 16C, each sub-image of the image is continuously written into the first image data region according to its high-frequency sub-band type. For example, taking sub-image 1 as an example, the coded macroblocks of the HL sub-band encoded data of sub-image 1 are continuously written into the first image data region, that is, sub-image 1-HL-MB0 to sub-image 1-HL-MBm of sub-image 1 are continuously written into the image bitstream. The coded macroblocks of the HH sub-band encoded data of sub-image 1 are continuously written into the first image data region, that is, sub-image 1-HH-MB0 to sub-image 1-HH-MBm of sub-image 1 are continuously written into the image bitstream. The coded macroblocks of the LH sub-band encoded data of sub-image 1 are continuously written into the first image data region, that is, sub-image 1-LH-MB0 to sub-image 1-LH-MBm of sub-image 1 are continuously written into the image bitstream. Other subgraphs are similar and will not be described in detail here. The order of subband types is for illustrative purposes only.

[0285] In this example, similar to the description in Figure 15, each coded subband of each sub-image can be decoded independently. Accordingly, offset information can be used to indicate the location of the image region to which the independently coded subband of each sub-image belongs. This allows, during decoding, the positions of the four coded subbands (LL subband encoded data, HH subband encoded data, HL subband encoded data, and LH subband encoded data) of a single sub-image can be obtained in the image data based on the offset information, and decoding can be performed on them to obtain the decoded subbands of the corresponding sub-image (including LL decoded data, HH decoded data, HL decoded data, and LH decoded data). Parts not described herein can be referred to Figure 15 and will not be repeated here.

[0286] For example, in the example of Figure 16C, during decoding, the decoder obtains the HH subband encoded data, LH subband encoded data, HL subband encoded data, and LL subband encoded data of sub-Figure 1 from the image bitstream based on offset information (not shown in Figure 16C, see Figure 15). The decoder decodes the encoded data of the four subbands of sub-Figure 1 to obtain the reconstructed sub-Figure 1, which can also be called decoded sub-Figure 1 or decoded sub-Figure 1.

[0287] In the embodiments of this application, multiple independently decoded encoded data can be decoded simultaneously, or one or more high-frequency subbands can be decoded simultaneously, and the number of simultaneous decodes depends on the decoder hardware performance.

[0288] Referring to Figure 17A, which is an exemplary schematic diagram of an image bitstream structure, the interleaving method in the example of Figure 17A is granular at the low-frequency and high-frequency subbands of the subgraph. Only the interleaving method of subgraph 1 is shown; other subgraphs are similar and will not be described in detail here. For example, the structure of the low-frequency subband encoded data of the subgraph can be seen in Figure 15, and will not be repeated here. In one example, the structure of the high-frequency subband encoded data of the subgraph can use any of the interleaving methods for high-frequency subband encoded data in Figures 16A to 16C.

[0289] Referring to Figure 17B, which is an exemplary schematic diagram of the bitstream structure, in the example of Figure 17B, the splicing order of the sub-graphs can be low-frequency subband coded data followed by high-frequency subband coded data. Other descriptions can be found in Figure 17A, and will not be repeated here.

[0290] Optionally, for the examples shown in Figures 17A and 17B, the offset information in the image header information is also used to indicate each encoded data that can be decoded independently. The indication method can be seen in Figures 15 and 16A to 16C, which will not be repeated here.

[0291] Referring to Figure 18, which is a schematic / conceptual block diagram of an exemplary decoder, in the example of Figure 18, the decoder receives, for example, an image bitstream encoded by an encoder to obtain a decoded image of the original image, also referred to as a decoded image, reconstructed image, etc. During the decoding process, the decoder receives the image bitstream from the encoder, including, but not limited to, image header information and image data. The image bitstream can be any of the bitstream formats shown in Figures 13A to 17B.

[0292] In the example shown in Figure 18, the decoder includes, but is not limited to: low-frequency subband processing path 1810, high-frequency subband processing path 1820, wavelet inverse transform unit 1830, image combination unit 1840 (also known as image stitching unit), etc.

[0293] For example, the low-frequency subband processing path 1810 is used to acquire low-frequency subband encoded data to obtain reconstructed low-frequency subband 1806 (also known as decoded low-frequency subband). The low-frequency subband processing path includes, but is not limited to: low-frequency subband entropy decoding unit 1811, inverse quantization / inverse transform unit 1812 (also known as low-frequency subband inverse quantization / inverse transform unit), low-frequency subband reconstruction unit 1813, low-frequency subband splicing unit 1815, prediction unit 1814, etc.

[0294] The high-frequency subband processing path 1820 is used to acquire high-frequency subband encoded data to obtain the reconstructed high-frequency subband 1831, which can also be called the reconstructed value of the high-frequency subband or the reconstructed data of the high-frequency subband, including but not limited to: high-frequency subband entropy decoding unit 1821, inverse quantization / inverse transform unit 1821 (also called high-frequency subband inverse quantization / inverse transform unit), high-frequency subband reconstruction unit 1831, etc.

[0295] In some instances, the decoder shown in Figure 18 can perform a decoding process that is largely the reverse of the encoding process described with reference to the encoder in Figure 12.

[0296] The decoding method in the embodiments of this application will be described in detail below with reference to the decoder 180 shown in Figure 18.

[0297] For example, the decoder 180 can obtain high-frequency subband encoded data and low-frequency subband encoded data in the image bitstream based on the image header information in the image bitstream. Furthermore, as described above, during the encoding process, the encoder uses macroblocks as the basic encoding unit, and correspondingly, during the decoding process, the decoder also uses macroblocks (e.g., encoded macroblocks) as the basic decoding unit for decoding.

[0298] For example, the low-frequency subband entropy decoding unit 1811 performs entropy decoding on the low-frequency subband encoded data 1801 in the image bitstream, using macroblocks as the basic decoding unit, to obtain the quantization coefficients 1802 (i.e., the quantization coefficients of the current macroblock) and syntax elements 1807 of the low-frequency subband block. The description of the quantization coefficients 1802 of the low-frequency subband can be found on the encoder side and will not be repeated here. Specifically, the low-frequency subband entropy decoding unit 1811 obtains the encoded macroblocks (i.e., the encoded data of the macroblocks) of the low-frequency subbands (e.g., LL subbands) of each sub-image in the image bitstream, and performs entropy decoding on each encoded macroblock to obtain the quantization coefficients 1802 (which can be simply referred to as the quantization coefficients of the macroblock of the low-frequency subband) and syntax elements 1807 of the corresponding low-frequency subband for each encoded macroblock. During the decoding process, the currently decoded encoded macroblock can be called the current block.

[0299] The low-frequency subband decoding unit is used to output the quantization coefficients 1802 of the low-frequency subband block to the inverse quantization / inverse transform unit 1812, and to output the syntax elements 1807 to the prediction unit 1814.

[0300] The inverse quantization / inverse transform unit 1812 is used to obtain the quantization coefficients 1802 of the low-frequency subband block to obtain the inverse quantization coefficients 1803 of the low-frequency subband block. Alternatively, it can be the inverse transform coefficients of the current block of the low-frequency subband (depending on whether inverse transform processing was performed). Specifically, the inverse quantization / inverse transform unit 1812 performs inverse quantization on the quantization coefficients of the current block of the low-frequency subband, or performs both inverse quantization and inverse transform, to obtain the inverse quantization coefficients of the current block of the low-frequency subband. The inverse quantization / inverse transform unit 1812 outputs the inverse quantization coefficients 1803 of the low-frequency subband block to the low-frequency subband reconstruction unit 1813, for example, the inverse quantization coefficients of the current block of the low-frequency subband.

[0301] The low-frequency subband reconstruction unit 1813 is used to obtain the low-frequency subband reconstruction block 1804, which can also be called the reconstruction coefficient of the low-frequency subband block, based on the quantization coefficients 1803 and prediction block 1805 of the low-frequency subband. Specifically, the low-frequency subband reconstruction unit 1813 adds a prediction block to the inverse quantization coefficients of the current block of the low-frequency subband to obtain the low-frequency subband reconstruction block 1804 corresponding to the current macroblock.

[0302] The prediction unit 1814 is used to obtain syntax element 1807 and perform corresponding prediction processing according to syntax element 1807. For example, intra-frame prediction can be performed based on low-frequency subband reconstruction block 1804, or inter-frame prediction can be performed based on reconstructed low-frequency subband 1806. Its execution method can be referred to the coding side, and will not be elaborated here. The prediction unit 1814 outputs prediction block 1805 to the low-frequency subband reconstruction block 1804 unit.

[0303] For example, the high-frequency subband entropy decoding unit 1821 acquires the high-frequency subband encoded data 1802 in the image bitstream, and, using macroblocks as the basic decoding unit, acquires the quantization coefficients 1808 (which are the quantization coefficients of the current macroblock) of each high-frequency subband block. Specifically, the high-frequency subband entropy decoding unit 1821 performs entropy decoding on the current block of the high-frequency subband encoded data 1802 to obtain the quantization coefficients of the current block of the high-frequency subband. Optionally, based on entropy decoding, the syntax elements corresponding to the current block can also be obtained. The high-frequency subband entropy decoding unit 1821 outputs the quantization coefficients 1808 of the high-frequency subband block to the inverse quantization / inverse transform unit 1822.

[0304] The inverse quantization / inverse transform unit 1822, also known as the high-frequency subband inverse quantization / inverse transform unit, is used to obtain the quantization coefficients 1808 of the high-frequency subband block to obtain the reconstruction coefficients 1809 of the high-frequency subband block. The reconstruction coefficients can be either inverse quantization coefficients after inverse quantization processing, or inverse transform coefficients after inverse quantization and inverse transform processing.

[0305] The high-frequency subband reconstruction unit 1823 (also known as the high-frequency subband splicing unit) is used to obtain the reconstruction coefficients 1809 of the high-frequency subband block to obtain the reconstructed high-frequency subband 1831, which can also be referred to as the reconstructed value or reconstructed data of the high-frequency subband. Specifically, the high-frequency subband reconstruction unit 1823 can obtain the reconstruction coefficients corresponding to each macroblock of the high-frequency subband, that is, reconstruct the high-frequency subband block. The high-frequency subband reconstruction unit 1823 can splice the obtained multiple macroblocks to obtain the corresponding high-frequency subband. Among them, the reconstructed high-frequency subband may optionally include reconstructing the HL subband (e.g., the reconstruction coefficients of the HL subband), reconstructing the HH subband (e.g., the reconstruction coefficients of the HH subband), and reconstructing the LH subband (e.g., the reconstruction coefficients of the LH subband).

[0306] The inverse wavelet transform unit 1830 is used to acquire the reconstructed high-frequency subband 1831 and the reconstructed low-frequency subband 1806 to obtain the reconstructed sub-image 1832. Specifically, the inverse wavelet transform unit 1830 acquires the reconstructed low-frequency subband 1806 output by the low-frequency subband stitching unit 1815 and the reconstructed high-frequency subband 1831 output by the high-frequency subband reconstruction unit 1823, and performs an inverse wavelet transform on the reconstructed low-frequency subband 1806 and the reconstructed high-frequency subband 1831 to obtain the reconstructed sub-image 1832. The inverse wavelet transform unit 1830 outputs the reconstructed sub-image 1832 to the image combining unit (which may also be called the image stitching unit, etc.).

[0307] Image combining unit 1840 is used to acquire reconstructed sub-images 1832 to obtain a reconstructed image 1833 of the original image, which can also be called a decoded image or a decoded image, etc. Specifically, image combining unit 1840 can acquire N reconstructed sub-images (N is an integer greater than 0) of the image (referring to the original image), and stitch (or combine) the N reconstructed sub-images according to the division method (including size and position) of each reconstructed sub-image during encoding to obtain the reconstructed image 1833.

[0308] Optionally, after obtaining the reconstructed image, the image combining unit 1840 can determine whether the reconstructed image contains a padding portion based on the image size information in the image header information and the size information of the current reconstructed image. In one example, if the size of the current reconstructed image is the same as the size indicated by the image size information (i.e., the same as the original image size), the image combining unit 1840 can send the reconstructed image to the display device. In this case, the sizes of the displayed image, the original image, and the reconstructed image are all the same. In another example, if the size of the current reconstructed image is different from the size indicated by the image size information (e.g., larger than the original image size), the image combining unit 1840 can remove the padding portion of the current reconstructed image based on the size indicated by the image size information to obtain the displayed image. The size of the displayed image is the same as the size of the original image. Specifically, in this embodiment, when the encoding side performs sub-image division, the division order is preset, usually from left to right in the horizontal direction and from top to bottom in the vertical direction. Correspondingly, at least one padding sub-image is usually located at the edge of the image, as shown in Figure 9B. For example, the image combining unit 1840 may crop the vertical and / or horizontal edges of the image according to the size information to remove the fill portion of the sub-image at the edge.

[0309] Optionally, in some instances, the image reconstruction unit may also perform the above-mentioned operation of removing the padding portion during the process of acquiring the reconstructed image, so that the size of the reconstructed image is the same as the size of the original image.

[0310] Optionally, the decoder is used, for example, to output a reconstructed image via the decoder's output port (or output interface) for presentation to or viewing by the user.

[0311] Other variations of the decoder can be used to decode compressed image bitstreams.

[0312] For example, as described above, each sub-image in the image bitstream may include two or four independently decodeable encoded data. For instance, in the examples shown in Figures 15, 16A, and 16C, the encoded HH sub-band, encoded LH sub-band, encoded HL sub-band, and encoded LL sub-band of each sub-image can be independently decoded. In the above scenario, the high-frequency sub-band entropy decoding unit can obtain each independently decodeable image data region based on the offset information in the image header information, and obtain the encoded data within the image data region. The high-frequency sub-band entropy decoding unit can simultaneously decode the encoded data of one or more independently decoded image data regions.

[0313] Optionally, in the above scenario, the high-frequency subband processing path may include at least one high-frequency processing subpath (not shown in the figure). For example, the high-frequency subband processing path may include three high-frequency processing subpaths to process three coded subbands of a subgraph simultaneously. Of course, in some instances, there may be more than three or fewer high-frequency subband processing subpaths. The more high-frequency subband processing paths there are, the higher the decoding efficiency. The fewer the paths, the lower the hardware design complexity requirements.

[0314] The decoding method is illustrated below using the bitstream structure shown in Figure 16A. In the example shown in Figure 16A, the low-frequency entropy decoding unit 1801 acquires the LL subband encoded data and performs decoding based on macroblocks. Taking sub-Figure 1 as an example, the low-frequency entropy decoding unit 1801 acquires the LL subband encoded data of sub-Figure 1 and decodes each encoded macroblock. The low-frequency subband decoding path 1810 processes the current block (inverse quantization / inverse transform, prediction, reconstruction, etc.) to output the reconstructed low-frequency subband of sub-Figure 1 to the wavelet inverse transform unit.

[0315] The high-frequency subband entropy decoding unit 1821 acquires LH subband encoded data, HL subband encoded data, and HH subband encoded data, and performs decoding based on the encoded macroblocks. Taking sub-image 1 as an example, specifically, the high-frequency entropy decoding unit 1821 acquires the LH subband encoded data, HL subband encoded data, and HH subband encoded data of sub-image 1 in the image bitstream based on offset information. The acquisition order can be in the order of the bitstream or according to actual needs. The high-frequency subband processing path 1820 processes the decoded macroblocks of sub-image 1 output by the high-frequency subband entropy decoding unit. In one example, the high-frequency subband entropy decoding unit can optionally decode the LH subband encoded data, HL subband encoded data, and HH subband encoded data of sub-Figure 1 simultaneously. The high-frequency subband processing path obtains multiple decoded high-frequency subbands output by the high-frequency subband entropy decoding unit and processes each macroblock in the multiple decoded high-frequency subbands simultaneously to output the reconstructed HH subband, reconstructed HL subband, and reconstructed LH subband of sub-Figure 1 to the wavelet inverse transform unit.

[0316] In another example, the high-frequency subband entropy decoding unit can optionally decode the LH, HL, and HH subband encoded data of sub-Figure 1 simultaneously. The high-frequency subband processing path acquires multiple decoded high-frequency subbands output by the high-frequency subband entropy encoding unit. It can process each macroblock in at least one decoded high-frequency subband, and the decoded data of other received but unprocessed high-frequency subbands can be cached in storage. The processing order can follow the order in the bitstream or be set according to actual needs. Similarly, the high-frequency subband processing path outputs to the wavelet inverse transform unit after acquiring a reconstructed high-frequency subband of sub-Figure 1.

[0317] After the wavelet inverse transform unit obtains the four reconstructed high-frequency subbands of sub-figure 1, including the reconstructed LL subband, reconstructed HL subband, reconstructed HH subband, and reconstructed LH subband, it performs wavelet inverse transform to obtain the reconstructed sub-figure 1.

[0318] Optionally, before acquiring the four reconstructed high-frequency subbands of sub-graph 1, the wavelet inverse transform unit can cache each reconstructed subband acquired. After acquiring all the reconstructed high-frequency subbands of sub-graph 1, it can acquire the cached reconstructed high-frequency subbands of sub-graph 1 and perform the wavelet inverse transform.

[0319] It can be understood that, in the embodiments of this application, when the high-frequency subband processing path and the low-frequency subband processing path process the encoded data of the sub-graph, the order of the sub-graphs and their macroblocks in each independent decoded data is the same. For example, as shown in Figure 15, the high-frequency subband entropy decoding unit can obtain the HH subband encoded data, HL subband encoded data, and LH subband encoded data of a single sub-graph (e.g., sub-graph 1) from the offset. In this way, when decoding, the high-frequency subband entropy decoding unit can obtain each encoded subband and its encoded macroblock of sub-graph 1 to decode sub-graph 1. Correspondingly, the high-frequency subband processing path can process the decoded macroblocks corresponding to the high-frequency subbands of sub-graph 1 to obtain each reconstructed high-frequency subband of sub-graph 1 and output it to the wavelet inverse transform unit so that the wavelet inverse transform unit can output the reconstructed sub-graph 1. In this way, the wavelet inverse transform unit only needs to buffer the decoded data of sub-graph 1 during the processing. If the high-frequency processing unit processes the encoded data in the code stream as shown in Figure 15, the wavelet inverse transform unit will cache the reconstructed HL subbands of other sub-graphs before obtaining the reconstructed HH subband of sub-graph 1, which increases the storage burden and requires a large hardware cache space, affecting the complexity of hardware design.

[0320] For example, in the example shown in Figure 16B, the high-frequency subband encoded data in the image bitstream is interleaved at the MB granularity. That is, during decoding, the high-frequency subband encoded data and low-frequency subband encoded data of each sub-image can be decoded independently. The high-frequency subband processing path can process each macroblock according to the order of the encoded macroblocks of the sub-image in the image bitstream, that is, the reconstructed subbands of each sub-image are obtained according to the order of the sub-images, which can reduce the hardware design complexity of the decoding end.

[0321] Referring to Figure 19, which is an exemplary schematic / conceptual block diagram of a decoder, the decoder 190 includes, but is not limited to: a low-frequency subband processing path 1910, a high-frequency subband processing path 1920, a wavelet inverse transform unit 1930, a sub-image combination unit 1940, and an image combination unit 1950. The low-frequency subband processing path 1910 includes, but is not limited to: a low-frequency subband entropy coding unit 1911, an inverse quantization / inverse transform unit 1912, a low-frequency subband reconstruction unit 1913, a low-frequency subband stitching unit 1915, and a prediction unit 1914. Detailed descriptions can be found in Figure 18, and will not be repeated here. The descriptions of the input and output coefficients or data of each unit (such as low-frequency subband coding data 1901, low-frequency subband quantization coefficients 1902, low-frequency subband inverse quantization coefficients 1903, low-frequency subband reconstruction block 1904, reconstructed low-frequency subband 1906, prediction block 1905, and syntax element 1907) can be found in Figure 18, and will not be repeated here.

[0322] The high-frequency subband processing path 1920 includes, but is not limited to: high-frequency subband entropy coding unit 1921, inverse quantization / inverse transform unit 1922, and high-frequency subband reconstruction unit 1923.

[0323] The high-frequency subband entropy coding unit 1921 obtains the quantization coefficients 1908 of the high-frequency subband block based on the high-frequency subband coded data 1902. The inverse quantization / inverse transform unit 1922 obtains the reconstruction coefficients 1909 of the high-frequency subband block based on the quantization coefficients 1908 of the high-frequency subband block, which can also be called the high-frequency subband reconstruction block.

[0324] The wavelet inverse transform unit 1903 is used to obtain the reconstruction coefficients 1909 of the high-frequency subband block, i.e., the high-frequency subband reconstruction block, and the low-frequency subband reconstruction block 1904 output by the low-frequency subband reconstruction block unit 1913. The wavelet inverse transform is performed on the high-frequency subband reconstruction block (e.g., including HH subband reconstruction block, HL subband reconstruction block, LH subband reconstruction block) and the low-frequency subband reconstruction block 1904 to obtain the reconstruction block 1931, which is the reconstruction block of the current subgraph. It can also be called the reconstruction data of the current block of the current subgraph or the reconstruction value of the current block of the current subgraph.

[0325] The wavelet inverse transform unit 1930 outputs a reconstructed block 1931 to the subgraph combination unit 1940. The subgraph combination unit 1940 can obtain the reconstructed subgraph of the current subgraph based on at least one reconstructed block corresponding to the current subgraph, which can also be referred to as the reconstructed value or reconstructed data of the current subgraph.

[0326] Subgraph combining unit 1940 outputs reconstructed subgraph 1932 to image combining unit 1950.

[0327] The image combination unit 1950 is used to acquire the reconstructed sub-image 1932 to obtain the reconstructed image 1933 of the original image, or the reconstructed value or reconstructed data of the original image, etc. The undescribed parts of Figure 19 can be referred to Figure 18, and will not be repeated here.

[0328] Figure 20a is a schematic diagram of the framework of an edge-cloud collaborative system provided in an embodiment of this application. The edge-cloud collaborative system may include: a central server, edge servers, and clients; wherein, a central server may connect to one or more edge servers, and an edge server may connect to one or more clients. This application does not limit the number of edge servers and clients; the specific number can be flexibly set according to the application scenario.

[0329] For example, a client can access the network through wireless access points such as base stations or Wi-Fi access points and communicate with the edge server through the network, or the client and the edge server can also communicate through a wired connection. Similarly, the edge server can also access the network through wireless access points such as base stations or Wi-Fi access points and communicate with the central server through the network, or the edge server and the central server can also communicate through a wired connection.

[0330] For example, the central server can be a single server, a server cluster consisting of multiple servers, or other distributed systems; this application does not impose any restrictions on this.

[0331] The client can be software, applications, browsers, in-vehicle systems, terminal devices, etc. When the client is implemented as a terminal device, it can include, but is not limited to, the following as shown in Figure 20a: mobile phone, personal computer (PC), virtual reality (VR) device, augmented reality (AR) device, tablet computer, laptop computer, etc. The client can also be a smart TV, mobile internet device (MID), wearable device (such as smartwatch, smart glasses, or smart helmet), smart car, wireless terminal device in industrial control, wireless terminal device in self-driving, wireless terminal device in remote medical surgery, wireless terminal device in smart grid, wireless terminal device in transportation safety, wireless terminal device in smart city, wireless terminal device in smart home, etc. The following embodiments do not impose special limitations on the specific form of the client.

[0332] In one possible scenario, some clients may connect directly to the central server instead of the edge server; in another possible scenario, all clients may connect directly to the central server instead of the edge server.

[0333] Furthermore, the edge-cloud collaborative system framework shown in Figure 20a is only an example of the edge-cloud collaborative system framework of this application. In the edge-cloud collaborative system of this application, the central server and the edge server can also be the same server; or the edge-cloud collaborative system of this application does not include edge servers, but the central server connects with each client. This application does not impose any restrictions on this.

[0334] As shown in Figure 20a, the end-to-cloud collaborative system can be applied to various image encoding and decoding scenarios of end-to-cloud collaboration, such as cloud gaming, cloud exhibitions, 3D cloud conferencing, 3D scenes, interior decoration, clothing design, architectural design and other multi-end collaborative image encoding and decoding scenarios. This application does not limit this.

[0335] In some embodiments, this application provides an encoding method and a decoding method. The functionality of the encoding method can be implemented by at least one of a server and a client, and the functionality of the decoding method can be implemented by at least one of a server and a client.

[0336] This server can be implemented through software or hardware.

[0337] In the first example, when the functionality of the server is implemented through software, the server may be, for example, an application running on a computing instance (such as an encoder or decoder, or an application that implements the encoding and decoding methods of this application), which may be, for example, a virtual machine, a container, or a host.

[0338] In the second example, when the server functionality is implemented through hardware, the server can be implemented through at least one physical device including a processor. This physical device can be a server (e.g., a central server or an edge server), a base station, a relay device, satellite equipment, etc., and there are no restrictions on this.

[0339] The processor can be a central processing unit (CPU) or a graphics processing unit (GPU), or it can be any type of processor or any combination thereof, such as an application-specific integrated circuit (ASIC), a programmable logic device (PLD), a complex programmable logical device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), a system-on-chip (SoC), a software-defined infrastructure (SDI) chip, an AI chip, or a data processing unit (DPU).

[0340] Furthermore, the number of processors included in the server can be arbitrary, and the types of processors included can be one or more. The specific number and types of processors can be set according to the actual business needs of the application, and this application does not impose any restrictions on this.

[0341] In the third example, when the server-side functionality is implemented through hardware, the server can also be a computing cluster comprising multiple computing nodes. Furthermore, these multiple computing nodes can communicate through the at least one switching node. Exemplarily, a computing node can be a computing server including an accelerator card. This accelerator card can be, for example, a deep-learning processing unit (DPU), a GPU, a neural-network processing unit (NPU), or a tensor processing unit (TPU), or other types of accelerator cards. Alternatively, a computing node can be a computing server including a general-purpose processor (such as a CPU).

[0342] The encoding method provided in this application can be applied to the wavelet transform-based encoding / decoding method and encoder / decoder of any of the above embodiments to encode and decode images and videos in any inter-frame prediction scenario.

[0343] The following example illustrates the encoding and decoding methods of this application by using an encoder to execute the encoding method and a decoder to execute the decoding method. However, this application does not limit the subject that executes the encoding and decoding methods of this application.

[0344] The encoding and decoding methods of this application will be illustrated below with different examples.

[0345] Example A1

[0346] Specific Example 1 in Example A1

[0347] In conjunction with any of the above embodiments, this application provides an encoding method, which may include the following steps:

[0348] Step 1: Obtain the reconstructed low-frequency subband of the reference image corresponding to the image to be encoded.

[0349] The image to be encoded includes blocks to be encoded, such as macroblocks.

[0350] When the image to be encoded is the original image, the reference image is also the original image encoded before the image to be encoded. When the image to be encoded is a sub-image, the reference image is the sub-image encoded before the image to be encoded. The image to be encoded and the reference image correspond to each other.

[0351] In related technologies, when performing inter-frame prediction, a reference image of the image to be encoded is used as a reference frame. Inter-frame prediction of the image to be encoded is performed by taking the pixel values ​​of pixels in the reference frame that are similar to those in the image to be encoded as the predicted values ​​of pixels in the image to be encoded, so as to obtain the prediction data for inter-frame prediction of the image to be encoded.

[0352] In the embodiments of this application, the reconstructed low-frequency subband of the reference image can be obtained.

[0353] The reconstructed low-frequency subband is obtained based on the wavelet coefficients of the low-frequency subband obtained by wavelet transforming the reference image.

[0354] In some embodiments, the encoder can perform wavelet transform on the reference image to obtain wavelet coefficients of the low-frequency sub-band; then, the reconstructed data of the wavelet coefficients of the low-frequency sub-band can be reconstructed based on the wavelet coefficients of the low-frequency sub-band, i.e., the low-frequency sub-band is reconstructed. The specific process can be seen in the process of obtaining the reconstructed low-frequency sub-band 1211 in Figure 12.

[0355] In some embodiments, the obtained reconstructed low-frequency subband may also be an adjusted reconstructed low-frequency subband (or a low-resolution map).

[0356] For example, the reconstructed low-frequency subband 1211 can be adjusted to obtain an adjusted reconstructed low-frequency subband.

[0357] For example, the low-frequency sub-band reconstruction block 1209 in Figure 12 can be spliced ​​to obtain the first reconstructed low-frequency sub-band, and then the first reconstructed low-frequency sub-band can be adjusted to obtain the adjusted reconstructed low-frequency sub-band. The reconstructed low-frequency sub-band 1211 is the adjusted reconstructed low-frequency sub-band.

[0358] In some embodiments, the adjusted reconstructed low-frequency subband can also serve as a reference submap.

[0359] In some embodiments, the adjustment may include range adjustment and / or precision adjustment. Range adjustment may involve reducing the offset to change the data range of the reconstructed low-frequency subband. Precision adjustment may involve shifting the data to the right by a certain number of bits (or bits per second) (e.g., shifting right by 2 bits, the specific number of bits is not limited) to reduce the data precision of the reconstructed low-frequency subband. The reconstructed low-frequency subband is used as a reference image to encode the current image to be encoded. By adjusting the reconstructed low-frequency subband, the amount of data in the reconstructed low-frequency subband can be reduced, thereby reducing memory usage and transmission overhead.

[0360] In some embodiments, range adjustment may be performed first, followed by precision adjustment of the range-adjusted data. In some embodiments, precision adjustment may be performed first, followed by range adjustment. In some embodiments, either range adjustment or precision adjustment may be performed.

[0361] In this embodiment, any image (e.g., a reference image) can be transformed by wavelet transform to obtain a high-frequency sub-band and a low-frequency sub-band. Both the high-frequency and low-frequency sub-bands can include wavelet coefficients. Therefore, in subsequent embodiments, the wavelet coefficients in the low-frequency sub-band are described as low-frequency sub-bands.

[0362] In the following text, considering that the low-frequency subband may contain wavelet coefficients, for the sake of brevity, the process of encoding and decoding the wavelet coefficients of the low-frequency subband will be described using the operation flow of encoding and decoding the low-frequency subband. The physical meaning of encoding and decoding the low-frequency subband and encoding and decoding the wavelet coefficients of the low-frequency subband are equivalent.

[0363] Step 2: Based on the reconstructed low-frequency subband, perform inter-frame prediction on the block to be coded to obtain the prediction data of the block to be coded (also known as the inter-frame prediction block).

[0364] In some embodiments, the prediction data for the block to be encoded is data from a reference block obtained from the reconstructed low-frequency subband.

[0365] For example, during inter-frame prediction, a reference block similar to the current block can be searched within the reconstructed low-frequency subband, and the data of that reference block in the reconstructed low-frequency subband can be used as the prediction block for the current block.

[0366] For example, if the current block size is 8x8, the reference block size is also 8x8. However, the search for a similar reference block within the reconstructed low-frequency subband is not based on each 8x8 block data segmented from the low-frequency subband of the reference image. Instead, it searches for an 8x8 region within the reconstructed low-frequency subband of the reference image based on pixel similarity, and uses this region as the reference block. The data from this reference block is then used as the predicted block for the current block.

[0367] In an optional embodiment, the encoder obtains a reference block from the reconstructed low-frequency subband, and based on this reference block, the encoder obtains the prediction data for the block to be encoded. For example, the encoder determines the data of the reference block in the reconstructed low-frequency subband as the prediction data for the block to be encoded.

[0368] In some embodiments, if the reconstructed low-frequency subband obtained in step 1 above is an adjusted reconstructed low-frequency subband, the adjusted reconstructed low-frequency subband can also be adjusted in the opposite way or partially in the opposite way.

[0369] For example, precision and / or range adjustments can be performed on the reconstructed low-frequency subband. For example, precision adjustment involves shifting left by a certain number of bits (or bit positions) (e.g., shifting left by 2 bits, the exact number of bits is not limited). Range adjustment involves increasing the offset.

[0370] For example, precision adjustment is performed on the reconstructed low-frequency subband, and then inter-frame prediction is performed on the block to be coded based on the precision-adjusted reconstructed low-frequency subband to obtain prediction data. Then, range adjustment is performed on the prediction data. This avoids range adjustment on all data in the reconstructed low-frequency subband, reducing the amount of computation.

[0371] In various embodiments of this application, both low-frequency subband and high-frequency subband may include wavelet coefficients. For the sake of brevity, the wavelet coefficients of the low-frequency subband are referred to as low-frequency subband, and the wavelet coefficients of the high-frequency subband are referred to as high-frequency subband.

[0372] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is 1 / 4 of the image before the wavelet transform (e.g., the image to be decoded), obtaining the reference block based on the reconstructed low-frequency subband reduces the amount of data acquired compared to obtaining the reference block based on the reference frame, thereby improving inter-frame prediction efficiency and coding efficiency. Considering that the data distribution of the reconstructed low-frequency subband is more uniform than that of the reconstructed high-frequency subband output by the wavelet transform, determining the data of the reconstructed low-frequency subband as the prediction data for the block to be coded results in more accurate prediction data, improving the accuracy of inter-frame prediction and thus improving coding accuracy.

[0373] This method can be implemented in conjunction with the wavelet transform-based encoder described above, as shown in the flowcharts of Figures 6, 8, and 12. In one example, step 2 can be implemented through the prediction unit 1237 shown in Figure 12 for inter-frame prediction. The specific principle can be found in the description of prediction unit 1237 above, and will not be repeated here.

[0374] Step 3: Based on the predicted data, encode the block to be encoded to obtain a bitstream.

[0375] For example, the actual value of the current block and the predicted value of the current block can be used to calculate the residual, and then the residual can be encoded to obtain the bitstream of the image to be encoded.

[0376] In some embodiments, the specific encoding process can refer to the process of encoding the current block (e.g., macroblock 1204) based on the prediction block as shown in Figures 6, 8, and 12 above, and will not be elaborated here.

[0377] For example, the image to be encoded may include multiple blocks of data, which can be encoded in the order of their encoding to obtain the bitstream described above.

[0378] Compared to acquiring a reference frame for inter-frame prediction of the current block within the current frame, this embodiment acquires the reconstructed low-frequency subband data (i.e., reconstructed low-frequency subband) of the reference frame to perform inter-frame prediction of the current block within the current frame. The size of the low-frequency subband of the reference frame is one-quarter of the reference frame's size, thus significantly reducing the amount of reference data acquired during inter-frame prediction, and consequently reducing the bit width occupied by the reference data. Since it is not necessary to read the entire image data of the reference frame, but only its reconstructed low-frequency subband, memory usage is reduced. Furthermore, performing inter-frame prediction of the current block based on the reconstructed low-frequency subband of the reference frame to obtain the predicted value of the current block for encoding the current frame improves the efficiency of inter-frame prediction compared to performing inter-frame prediction based on the reference frame, thereby improving encoding efficiency.

[0379] Based on any of the above wavelet transform-based embodiments (e.g., embodiments of any encoding method in Figures 6 to 19) and any of the encoding methods involving inter-frame prediction in Figures 1 to 5, the encoding methods of the embodiments of this application will be described below with reference to different examples.

[0380] Specific Example 2 in Example A1

[0381] Figure 20b illustrates a flowchart of an encoding method according to this application. As shown in Figure 20b, the method flow may include, but is not limited to, the following steps:

[0382] S101. Perform wavelet transform on the current frame to obtain the low-frequency sub-band of the current frame.

[0383] Optionally, S102. Divide the low-frequency subband of the current frame into blocks to obtain the current block.

[0384] The current frame can be either a video frame or an image frame; there are no restrictions.

[0385] The current frame can also be a sub-image in a video or a sub-image in an image, without any restrictions.

[0386] In an optional embodiment, the current block is the block data in the wavelet coefficients of the low-frequency subband obtained by wavelet transform of the current frame.

[0387] In a specific embodiment, after the encoder performs wavelet transform on the current frame, it can obtain a low-frequency sub-band. The encoder can divide the low-frequency sub-band into blocks to obtain multiple block data, where the current block is a block data to be encoded.

[0388] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is 1 / 4 of the image before the wavelet transform (e.g., the image to be encoded), performing inter-frame prediction on the block data in the low-frequency subband reduces the amount of data encoded compared to performing inter-frame prediction and encoding on the image to be encoded, thereby improving the efficiency of inter-frame prediction and encoding. Considering that the data distribution in the low-frequency subband is more uniform than that in the high-frequency subband output by the wavelet transform, performing inter-frame prediction and encoding on the block data in the low-frequency subband using the method of this embodiment can improve the accuracy of inter-frame prediction, thereby improving the encoding accuracy. Furthermore, it allows the encoded block data of the low-frequency subband to be used for display, for example, the displayed image of the low-frequency subband can be a thumbnail of the image to be encoded (e.g., the original image to be encoded).

[0389] In an optional embodiment, the encoder divides the low-frequency subband into multiple data blocks, and the encoder determines the current block from these multiple data blocks. Of course, this application does not limit the block division method of the image to be encoded, nor the number of blocks obtained.

[0390] In some embodiments, each block is 8x8 in size. In other embodiments, the block size may also be 8x4 or 4x4, and there is no limitation herein.

[0391] To illustrate with a specific example 1-1, please refer to Figure 21a. Figure 21a(1) is a schematic diagram of the low-frequency sub-band of the current frame provided in an embodiment of this application. The encoder divides the low-frequency sub-band of the current frame into 9 data blocks, which are arranged in the encoding order as block 1->block 2->block 3->block 4->block 5->block 6->block 7>block 8->block 9. Similarly, the encoding order and the decoding order can be the same, so the decoder can also decode each block in the order of block 1->block 2->block 3->block 4->block 5->block 6->block 7>block 8->block 9. For example, the current block on the encoding side can be block 5. Figure 21a(2) is a schematic diagram of the reconstructed low-frequency sub-band of the reference frame corresponding to the current frame provided in an embodiment of this application. The encoder divides the low-frequency sub-band of the reference frame into 9 data blocks, which are arranged in the encoding order as block 1'->block 2'->block 3'->block 4'->block 5'->block 6'->block 7'>block 8'->block 9'. Similarly, the encoding order and the decoding order can be the same, so the decoder can also decode each block in the following order: block 1'-> block 2'-> block 3'-> block 4'-> block 5'-> block 6'-> block 7'> block 8'-> block 9'.

[0392] S103a. Obtain the reconstructed low-frequency subband of the reference frame.

[0393] The specific implementation principle of S103a is the same as that of step 1 in the above specific example 1, and will not be repeated here.

[0394] This application does not impose any restrictions on the execution order of S101, S102, and S103a.

[0395] Optionally, S103b. Get the actual value of the current block.

[0396] In an optional embodiment, the encoder stores the actual value of the current block, which is the data to be encoded for each pixel in the current block. This application does not limit this.

[0397] In some embodiments, the actual value of the current block is the block data (e.g., wavelet coefficients) of the current block in the low-frequency subband of the current frame obtained in S102 of the current frame as shown in FIG20b.

[0398] This application embodiment does not restrict the execution order between S103a and S103b.

[0399] S104. Based on the reconstructed low-frequency subband of the reference frame, perform inter-frame prediction on the current block to obtain the predicted value and MVD of the current block.

[0400] The relevant description of S104 can be found in the relevant content of Specific Example 1 above. S105. Determine the residual of the current block based on the actual value and the predicted value of the current block.

[0401] In an optional embodiment, the encoder determines the residual of the current block as the difference between the actual value of the current block and the predicted value of the current block.

[0402] S106. Encode based on the residual of the current block and the MVD of the current block to obtain the bitstream.

[0403] In an optional embodiment, the encoder encodes the residual of the current block and the MVD of the current block to obtain a bitstream.

[0404] This application does not limit the specific encoding process of the MVD of the current block and the predicted value of the current block obtained by inter-frame prediction. It can be the encoding implementation process of the inter-frame prediction result of any of the above embodiments. This application also does not limit the syntax elements in the bitstream. Specifically, it can be combined with the bitstream obtained by the encoding method or encoder involving wavelet transform in any of the above embodiments.

[0405] In an optional embodiment, the specific implementation of S104 shown in FIG20b can be achieved through the process shown in FIG22. As shown in FIG22, the process may include, but is not limited to, the following steps:

[0406] S141. Based on the data of the reconstructed low-frequency subband of the reference frame and the current block in the low-frequency subband of the current frame, obtain the reference block in the reconstructed low-frequency subband of the reference frame.

[0407] In an optional embodiment, the encoder searches for a reference block similar to the current block data from the reconstructed low-frequency subband of the reference frame based on the data of the current block, and the encoder obtains the reference block similar to the current block data.

[0408] In some embodiments, the encoder also obtains the reconstructed value of the reference block based on the reconstructed low-frequency subband of the reference frame.

[0409] S142. Based on the reference frame, reconstruct the position of the reference block and the current block within the low-frequency subband, and obtain the MV of the current block.

[0410] In an optional embodiment, the encoder obtains the position of the reference block in the reconstructed low-frequency subband of the reference frame, and calculates the MV of the current block based on the position of the reference block in the reconstructed low-frequency subband of the reference frame and the position of the current block.

[0411] S143. Based on the MV of the current block and the MV of the coded blocks in the current frame, obtain the MVD of the current block.

[0412] In an optional embodiment, the block that has been encoded in the current frame is the block that is adjacent to the current block in the current frame.

[0413] In an optional embodiment, the encoder has already calculated the MV of the encoded blocks within the current frame before calculating the MV of the current block. The encoder stores the MV of the encoded blocks within the current frame and retrieves the MV of the encoded blocks within the current frame. The encoder determines the MVD of the current block as the difference between the MV of the current block and the MV of the encoded blocks within the current frame.

[0414] S144. Reconstruct the reference block within the low-frequency subband based on the reference frame, and obtain the prediction value of the current block.

[0415] In an optional embodiment, the predicted value of the current block is the reconstructed value of the reference block obtained from the reconstructed low-frequency subband of the reference frame.

[0416] In a specific embodiment, the encoder obtains a reference block from the reconstructed low-frequency subband of the reference frame, and based on this reference block, the encoder obtains the predicted value of the current block. For example, the encoder determines the reconstructed value of the reference block as the predicted value of the current block.

[0417] For example, the encoder in this embodiment may include a prediction unit 1237 as shown in FIG12 to perform the inter-frame prediction process shown in FIG22. Specifically, the prediction unit 1237 may search for a reference block in the reconstructed low-frequency subband of the reference frame that is similar to the data of the current block, and use the reference block as the prediction block of the current block. The reconstructed low-frequency subband is input to the prediction unit 1237 for inter-frame prediction to reduce the memory usage of the prediction unit 1237.

[0418] Furthermore, as shown in Figure 22, the prediction unit 1237 can obtain the motion vector (MV) of the current block based on the position of the current block, the searched reference block, and its position in the reconstructed low-frequency subband of the reference frame.

[0419] Then, the prediction unit 1237 can obtain the MVD of the current block based on the MV of the current block and the MV of the encoded blocks in the current frame (which are the blocks adjacent to the current block in the current frame).

[0420] In some embodiments, the current frame (or the data of the current block in the low-frequency subband) shown in FIG22 is subjected to wavelet transform (also known as wavelet forward transform) and adjustment operations.

[0421] In the process shown in Figure 22, the method can further perform the above-mentioned adjustment operation on the reconstructed low-frequency sub-band of the reference frame, and then, based on the reconstructed low-frequency sub-band after the adjustment operation, perform the inter-frame prediction process shown in Figure 22 on the current block after wavelet transform and adjustment operation to obtain the prediction value of the current block.

[0422] Since the reference block in the reconstructed low-frequency subband obtained in S141 is searched from the reconstructed low-frequency subband after the adjustment operation, the data of the reference block obtained in S141 has also been adjusted, thus making the predicted value of the current block obtained in S144 equivalent to having been adjusted.

[0423] In some embodiments, the above-described adjustment operation may include at least one of precision adjustment based on a first adjustment parameter and range adjustment based on a second adjustment parameter.

[0424] For example, precision adjustment can be achieved by shifting the data to the left by a certain number of bits (or bits per second) (e.g., shifting left by 2 bits, the specific number of bits is not limited) to improve the precision of the data being adjusted. Range adjustment can be achieved by increasing the offset to change the data range of the data being adjusted. The offset can be a positive or negative number, without limitation.

[0425] In some embodiments, the first adjustment parameter may include an adjustment amount and an adjustment type. For example, if the first adjustment parameter is a left shift of 2 bits, then the adjustment type is a left shift and the adjustment amount is 2 bits.

[0426] In some embodiments, the second adjustment parameter may include an adjustment amount, an adjustment type, and an adjustment direction; for example, the second adjustment parameter is to increase the offset. The adjustment type is offset, the adjustment direction is increase, and the adjustment amount is... For example, the second adjustment parameter is to increase the offset. The adjustment type is offset, the adjustment direction is decrease, and the adjustment amount is...

[0427] Thus, the prediction block of the current block is a reference block searched from the reconstructed low-frequency subband of the reference frame after precision adjustment and / or range adjustment, so that the prediction block is precision adjusted and / or range adjusted.

[0428] In some embodiments, the predicted value of the current block is data of a reference block obtained from a first region of the reconstructed low-frequency subband of a reference frame. This first region is obtained by expanding a specified number of samples around the corresponding position of the current block in the reconstructed low-frequency subband (hereinafter referred to as the "reference position").

[0429] In an optional embodiment, the encoder expands the corresponding position of the current block in the reconstructed low-frequency subband by a specified number of samples to obtain a first region. This first region is the region determined in the reconstructed low-frequency subband. Then, the encoder can search for a reference block similar to the current block from the first region of the reconstructed low-frequency subband of the reference frame, and use the data of the reference block located in the first region within the reconstructed low-frequency subband as the predicted value of the current block.

[0430] The reference block data is the data of an 8x8 block in the reconstructed low-frequency subband. Specifically, the encoder can search for an 8x8 block with a pixel value close to that of the current block within the first region of the reconstructed low-frequency subband to obtain the reference block data.

[0431] For example, the encoder shown in FIG12 can implement the encoding method of the present application embodiment. Specifically, the prediction unit 1237 shown in FIG12 can perform inter-frame prediction on macroblock 1204 (an example of the current block) based on the reconstructed low-frequency subband 1211 of the reference image to obtain prediction block 1205.

[0432] The specific process of inter-frame prediction can be implemented by prediction unit 1237 as shown in Figure 12 according to the inter-frame prediction process, and this application does not impose any restrictions.

[0433] For example, prediction unit 1237 includes an inter-frame prediction unit (not shown), which may include a motion estimation (ME) unit (not shown in FIG12) and a motion compensation (MC) unit (not shown in FIG12). The functions of the motion estimation unit and the motion compensation unit can be referred to the description of the above embodiments.

[0434] In a specific embodiment, under the same coordinate system, the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame is the same as the position of the current block in the low-frequency subband of the current frame.

[0435] In an optional embodiment, the encoder may determine the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame based on the position of the current block in the low-frequency subband of the current frame (e.g., the low-frequency subband obtained by wavelet transform of the current frame).

[0436] Referring to the specific example 1-1 above, and referring to Figure 21a, the size of each block is 8x8. Referring to Figure 21a(1), the current block in the low-frequency sub-band of the current frame is block 5. For example, the corresponding position of block 5 in the reconstructed low-frequency sub-band of the reference frame shown in Figure 21a(2) (i.e., the above-mentioned reference position) is the position of block 5' in the reconstructed low-frequency sub-band of the reference frame.

[0437] In a specific embodiment, the encoder can expand a specified number of sample points around the position of block 5' as shown in Figure 21a(2) to obtain the first region 10.

[0438] As shown in Figure 21a(2), the surrounding area may include at least one side of the horizontal direction from the position of block 5', and / or at least one side of the vertical direction from the position of block 5'. Thus, the encoder expands to at least one side of the horizontal left and right sides of the position of block 5' (e.g., a plurality of arrows on the left and right sides of block 5', expanding the left and right sides in the direction of the arrows) and to at least one side of the vertical up and down sides of the position of block 5' (e.g., a plurality of arrows on the up and down sides of block 5', expanding the up and down sides in the direction of the arrows) to obtain the first region 10 shown in the dashed box.

[0439] Of course, the number of sample points expanded in the four directions of up, down, left, and right can be the same or different, and there is no restriction here.

[0440] Furthermore, the shape of the first region is not limited to a rectangle; it can also be an irregular shape.

[0441] Then, as shown in Figure 21a(2), the encoder can search within the first region 10 for an 8x8 region with a pixel value similar to that of block 5 (example of the current block) shown in Figure 21a(1), so that the block data of the 8x8 region in the reconstructed low-frequency subband of the reference frame (i.e., the data of the reference block) is used as the prediction block of block 5 in the inter-frame prediction of the reference frame shown in Figure 21a(1).

[0442] In this way, when the encoder performs inter-frame prediction on the current block within the current frame based on the reconstructed low-frequency subband of the reference frame, it can obtain the reference block data within the first region of the reconstructed low-frequency subband, which can then be used as the prediction data for the inter-frame prediction of the current block. Compared to searching for a reference block similar to the current block within the reconstructed low-frequency subband of the reference frame, the present embodiment searches within the first region, which narrows the range of reference blocks, ensuring that the reference blocks are not far from the current block and remain within the controllable first region. This reduces the bit width of the current block's MV, thereby improving the prediction efficiency of inter-frame prediction and reducing the amount of encoded data to improve coding efficiency. Furthermore, the first region is obtained by expanding the area around the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame, ensuring that the region is still near the corresponding position of the current block. Thus, the reference blocks searched for in inter-frame prediction are also quite similar to the current block, ensuring the accuracy of inter-frame prediction, thereby reducing the amount of encoded data and improving coding efficiency. In this embodiment of the application, inter-frame prediction is performed in the first region to obtain the prediction data of the current block. Compared with determining the prediction data of the current block in the entire reference frame, the range of the MV of the current block can be narrowed, so that the MV of the current block is fixed within a certain range.

[0443] Furthermore, by using a fixed first region as the search reference block, the final determined range of MV is also fixed. Therefore, a fixed bit width can be used to represent MV, eliminating the need to set additional bit widths to avoid incomplete representation of MV. This saves bit width, reduces the amount of data encoded, and improves encoding efficiency. For example, MV can be represented using 4 bits.

[0444] In an optional embodiment, the first region is a region obtained by expanding the same number of first reference points to both sides in the horizontal direction based on the above-mentioned "reference position".

[0445] In a specific embodiment, the encoder expands the first number of points horizontally to the left and right from the "reference position" to obtain the first region. For example, the first number of points is 7. In other embodiments, the first number of points can be any integer greater than 0, and can be flexibly set according to needs without limitation.

[0446] The embodiments of this application do not limit the order in which the encoder expands to both sides in the horizontal direction.

[0447] In this embodiment, since the first region is the search range for searching the MV of the current block, it is set as a region centered on the corresponding position of the current block in the reconstructed low-frequency subband, expanding the same number of samples to the left and right sides in the horizontal direction. This ensures the accuracy of the searched MVs in scenarios where the MV of the current block is searched with equal probability on both sides of the corresponding position in the horizontal direction.

[0448] In other embodiments, in scenarios where prior information or other means determine that the MV of the current block will be searched with different probabilities on either side of the corresponding position (i.e., the aforementioned "reference position") in the horizontal direction, for example as shown in Figure 21a(2), compared to the left side of block 5' in the horizontal direction, the reference block of block 5 has a higher probability of appearing on the left side of block 5' in the horizontal direction. Then, the first region can be a region obtained by expanding the number of sample points to the left by L and the number of sample points to the right by R in the horizontal direction based on the aforementioned "reference position", where R is greater than L, and R and L are both positive integers. In this way, since the MV of the current block is more likely to be searched on one side of the aforementioned "left" and "right" sides of the "reference position", more sample points can be expanded on that "side".

[0449] In an optional embodiment, the first region is a region obtained by expanding the same number of second sample points to both sides in the vertical direction based on the above-mentioned "reference position".

[0450] In a specific embodiment, the encoder expands the number of second sample points vertically upwards and downwards from the "reference position" to obtain the first region. For example, the number of second sample points is 3. In other embodiments, the number of second sample points can be any integer greater than 0, and can be flexibly set according to requirements without limitation.

[0451] This application does not limit the order in which the encoder expands to both sides in the vertical direction. In a specific embodiment, the encoder expands the second number of sample points downwards in the vertical direction from the "reference position," and then expands the second number of sample points upwards in the vertical direction to obtain the first region. For example, the second number of sample points is 3.

[0452] In an optional embodiment, the first region is a region obtained by expanding the number of sample points upward in the vertical direction by D1 and the number of sample points downward by D2 based on the "reference position".

[0453] In a specific embodiment, the encoder expands the number of sample points D1 vertically upward from the "reference position" and then expands the number of sample points D2 vertically downward to obtain the first region.

[0454] In a specific embodiment, after the encoder expands the number of sample points D1 downward in the vertical direction from the "reference position", it expands the number of sample points D2 upward in the vertical direction to obtain the first region.

[0455] It should be noted that the embodiments of this application do not impose restrictions on the specific values ​​of the number of sample points D1 and D2, as long as both the number of sample points D1 and D2 are positive integers greater than or equal to 0.

[0456] It should be noted that the embodiments of this application do not limit the size relationship between the number of sample points D1 and the number of sample points D2. That is, the embodiments of this application do not limit the size relationship between the range of the area expanded upward by the encoder and the range of the area expanded downward by the encoder.

[0457] In one example, the number of samples D1 can be greater than the number of samples D2. For example, the number of samples D1 can be 4 and the number of samples D2 can be 2. In another example, the number of samples D1 can also be less than the number of samples D2. For example, the number of samples D1 can be 1 and the number of samples D2 can be 5.

[0458] In an optional embodiment, the encoder first expands the horizontal direction by a specified number of samples from the "reference position," and then expands the vertical direction by a specified number of samples to obtain the first region. Alternatively, the encoder first expands the vertical direction by a specified number of samples from the "reference position," and then expands the horizontal direction by a specified number of samples to obtain the first region.

[0459] This application does not limit whether the specified number of samples expanded upwards by the encoder is the same as the specified number of samples expanded downwards by the encoder; nor does it limit whether the specified number of samples expanded to the left by the encoder is the same as the specified number of samples expanded to the right by the encoder. In other words, this application does not limit whether the range of the encoder's upward expansion is equal to the range of the encoder's downward expansion; nor does it limit whether the range of the encoder's left expansion is equal to the range of the encoder's right expansion.

[0460] In some embodiments, the specified number of samples expanded upwards by the encoder can be the same as the specified number of samples expanded downwards by the encoder, and the specified number of samples expanded to the left by the encoder can be the same as the specified number of samples expanded to the right by the encoder. For example, the encoder expands 3 samples to both the upward and downward sides, and expands 7 samples to both the left and right sides.

[0461] It should be noted that the embodiments of this application do not limit the relationship between the specified number of samples expanded in the horizontal direction and the specified number of samples expanded in the vertical direction. That is, the embodiments of this application do not limit the relationship between the range of the encoder's horizontally expanded area and the range of the encoder's vertically expanded area.

[0462] In some embodiments, the first sample number is greater than the second sample number. For example, the first region determined by the encoder is the region located in the reconstructed low-frequency subband obtained by expanding the number of samples in the horizontal direction with the "reference position" as the center, which is greater than the number of samples expanded in the vertical direction.

[0463] Referring to Figure 21a above, and then to Figure 21b(1), this application embodiment provides a schematic diagram of the first region within the reconstructed low-frequency subband of another reference frame.

[0464] Figure 21a(2) and Figure 21b(1) show the reconstructed low-frequency subband of the same reference frame. The position of block 5' is the corresponding position of block 5 in the reconstructed low-frequency subband as shown in Figure 21a(1) (i.e., the aforementioned "reference position"). The size of the current block (block 5) is 8x8, and the size of block 5' in the reconstructed low-frequency subband as shown in Figure 21b(1) is also 8x8.

[0465] Figure 21b(1) shows the first region 11 obtained by expanding the first number of sample points in the horizontal direction (specifically, 7 sample points on each of the left and right sides, with the first number of sample points being 7) and expanding the second number of sample points in the vertical direction (specifically, 3 sample points on each of the top and bottom sides, with the second number of sample points being 3) based on the reference position where block 5' is located.

[0466] Thus, the number of samples expanded horizontally is greater than the number of samples expanded vertically. In this embodiment, considering that the encoder reads data in the reconstructed low-frequency subband of the reference frame line by line from the hardware, if the number of samples is greater than the number of samples, the encoder will read fewer lines of data when reconstructing the data in the first region of the low-frequency subband. Therefore, the number of times line data needs to be read increases, requiring more line buffers and thus consuming too much memory. Therefore, in this embodiment, the specified number of samples expanded horizontally in the first region is greater than the specified number of samples expanded vertically by the encoder. This reduces the search range of the MV while also reducing the number of times the encoder reads line data in the first region, thereby reducing memory consumption during inter-frame prediction.

[0467] In the embodiment of Figure 21a(1), the size of the block is 8x8, and it is expanded by 7 samples on both sides in the horizontal direction and by 3 samples on both sides in the vertical direction to obtain the first region. The first region obtained in this way can be the region with the highest probability of the reference block appearing in the inter-frame prediction of the current block. Then, inter-frame prediction is performed in the first region to obtain the prediction block of the current block, which can improve the prediction efficiency and prediction accuracy of the current block, thereby improving the coding speed and coding efficiency.

[0468] Based on the first region 11 obtained in Figure 21b(1), please refer to Figure 21b(2). The encoder can search for a reference block 20 similar to block 5 shown in Figure 21a(1) within the first region 11, so as to use the data of the reference block 20 (size 8x8) in the reconstructed low-frequency subband shown in Figure 21b(1) as the prediction block of block 5, thereby obtaining the MV of block 5.

[0469] In Figures 21a and 21b(1) above, taking block 5 as shown in Figure 21a(1) as an example, the location of the first region is illustrated.

[0470] In another example, referring to Figure 21b(3) above, in conjunction with Figure 21a above, this application embodiment provides a schematic diagram of the first region within the reconstructed low-frequency subband of another reference frame.

[0471] Unlike Figure 21b(1), in Figure 21b(3), the current block is block 1 as shown in Figure 21a(1). Figures 21a(2) and 21b(3) show the reconstructed low-frequency subband of the same reference frame. The position of block 1' is the corresponding position of block 1 in the reconstructed low-frequency subband as shown in Figure 21a(1) (i.e., the aforementioned "reference position"). The size of the current block (block 1) is 8x8, and the size of block 1' in the reconstructed low-frequency subband as shown in Figure 21b(3) is also 8x8.

[0472] Figure 21b(3) shows the first region 12, which is the region obtained by expanding the first number of sample points in the horizontal direction (specifically, 7 sample points on each of the left and right sides, with the first number of sample points being 7) and expanding the second number of sample points in the vertical direction (specifically, 3 sample points on each of the top and bottom sides, with the second number of sample points being 3) based on the reference position where block 1' is located.

[0473] As shown in Figure 21b(3), during the process of the encoder expanding the first number of sample points (e.g., 7 sample points) to the left and right sides in the horizontal direction and expanding the second number of sample points to the upper and lower sides in the vertical direction, the expanded sample points are not in the reconstructed low-frequency sub-band because the position of block 1' in the reconstructed low-frequency sub-band is located in the upper left corner.

[0474] For example, regions 1 to 5, as shown in Figure 21b(3), are not within the reconstructed low-frequency subband shown in Figure 21b(3). Therefore, the encoder can pad regions 1 to 5 based on the data in the reconstructed low-frequency subband shown in Figure 21b(3) to obtain the complete data for the first region 12. Then, the data of the reference block can be searched within the first region 12 to serve as the prediction block for the inter-frame prediction of the current block (here, block 1).

[0475] When padding regions 1 to 5, block 1' is adjacent to regions 2, 4, 6, and 7, respectively. The encoder can then copy the first row of data from block 1' three times to create three rows, thus filling region 2 with data where each row in region 2 is identical to the first row of data from block 1'. Similarly, it can copy the first column of data from block 1' seven times to create seven columns, filling region 4 with data where each column in region 4 is identical to the first column of data from block 1'.

[0476] Following a similar approach to padding regions 2 and 4, the encoder can pad region 1 based on the data in the first column of region 2 adjacent to region 1 and the data in the first row of region 4 adjacent to region 1; similarly, the encoder can pad region 3 based on the data in the first column of region 2 adjacent to region 3 and the data in the first row of region 6 adjacent to region 3; similarly, the encoder can pad region 5 based on the data in the first row of region 4 adjacent to region 5 and the data in the first column of region 7 adjacent to region 5.

[0477] Then, data of a reference block similar to block 1 can be obtained in the first region 12 shown in Figure 21b(3) as the prediction data for the current block.

[0478] Thus, when the reconstructed low-frequency subband of the reference frame cannot cover the first region, reference data within the first region (i.e., data located in the first region within the reconstructed low-frequency subband, such as data in region 12 shown in FIG21b(3)) can be obtained by filling data around the "reference position" based on the data of the reconstructed low-frequency subband. Obviously, the reference data in region 12 shown in FIG21b(3) is also obtained based on the reconstructed low-frequency subband. Therefore, when the reconstructed low-frequency subband of the reference frame cannot cover the first region, since the additional data filled in the first region is based on the data of the reconstructed low-frequency subband, the first region in each embodiment of this application can also be described as the first region of the reconstructed low-frequency subband (or described as the first region within the reconstructed low-frequency subband). Similarly, when the reconstructed low-frequency subband of the reference frame cannot cover the first region, since the additional data filled in the first region is based on the data of the reconstructed low-frequency subband, the prediction data of the current block obtained by searching within the first region can also be described as the data of the reference block obtained from the reconstructed low-frequency subband.

[0479] In some embodiments, the first region is a region comprising a reconstructed low-frequency subband and expanded around the reconstructed low-frequency subband by a specified number of samples. For example, as shown in FIG21c, the reconstructed low-frequency subband includes multiple blocks, such as block 1', block 2', etc. The reconstructed low-frequency subband is expanded in the horizontal and / or vertical directions to obtain the first region 13.

[0480] When performing motion search in the reconstructed low-frequency subband, if the search range of the reference block is at the edge of the reconstructed low-frequency subband, the expanded reconstructed low-frequency subband is used for the search, which improves the accuracy of the motion search and reduces the computational complexity of the motion search.

[0481] In some embodiments, the reconstructed low-frequency subband is extended in the horizontal direction, for example, by extending it to the left and / or right side in the horizontal direction of the reconstructed low-frequency subband.

[0482] In some embodiments, the reconstructed low-frequency subband is extended in the vertical direction, for example, by extending it on the upper and / or lower side of the reconstructed low-frequency subband in the vertical direction.

[0483] It should be understood that when expanding the reconstructed low-frequency subband, the number of samples to be expanded in the horizontal and / or vertical directions can be referred to the relevant description in Figure 21b.

[0484] In some embodiments, when expanding the reconstructed low-frequency subband, expansion is performed on both the luminance and chrominance components. For example, the number of samples expanded horizontally for the luminance component is not less than 7, and the number of samples expanded vertically is not less than 3. For example, the number of samples expanded horizontally for the chrominance component is not less than (7 + FormatShiftX) >> FormatShiftX), where FormatShiftX is 0 or 1, and the number of samples expanded vertically is not less than 3.

[0485] In some embodiments, when expanding the reconstructed low-frequency subband, the expanded sample can be replaced or expanded using the nearest integer sample (edge ​​sample or corner sample) within the reconstructed low-frequency subband. That is, the value (wavelet coefficient) of the expanded sample is consistent with the value of the nearest integer sample within the reconstructed low-frequency subband.

[0486] In some embodiments, when the codec hardware performance is strong, the number of samples added horizontally can be less than the number of samples added vertically by the encoder. For example, the number of samples added horizontally is 10, and the number of samples added vertically is 12. That is, the number of samples added to the left by the encoder is 5, the number of samples added to the right by the encoder is 5, the number of samples added to the upper side by the encoder is 6, and the number of samples added to the lower side by the encoder is 6.

[0487] In another embodiment, the specified number of samples added by the encoder in the horizontal direction can be equal to the specified number of samples added by the encoder in the vertical direction. For example, the number of samples added in the horizontal direction is 10, and the number of samples added in the vertical direction is 10. That is, the number of samples added by the encoder to the left, the number of samples added by the encoder to the right, the number of samples added by the encoder to the upper side, and the number of samples added by the encoder to the lower side are all 5.

[0488] In other embodiments, the specified number of samples expanded upwards by the encoder can be the same as the specified number of samples expanded downwards by the encoder, and the specified number of samples expanded to the left by the encoder can be different from the specified number of samples expanded to the right by the encoder. For example, the encoder expands 3 samples both upwards and downwards, expands 8 samples to the left by the encoder, and expands 6 samples to the right by the encoder.

[0489] In some other embodiments, the specified number of samples that the encoder expands upwards may be different from the specified number of samples that the encoder expands downwards, and the specified number of samples that the encoder expands to the left may be the same as the specified number of samples that the encoder expands to the right. For example, the encoder expands upwards by 4 samples, expands downwards by 2 samples, and expands to the left and right by 7 samples each.

[0490] In some other embodiments, the specified number of samples that the encoder expands upwards may be different from the specified number of samples that the encoder expands downwards, and the specified number of samples that the encoder expands to the left may be different from the specified number of samples that the encoder expands to the right. For example, the encoder expands upwards by 4 samples, expands downwards by 2 samples, expands to the left by 8 samples, and expands to the right by 6 samples.

[0491] It should be understood that when the specified number of samples expanded upwards by the encoder differs from the specified number of samples expanded downwards, and / or when the specified number of samples expanded to the left differs from the specified number of samples expanded to the right, the embodiments of this application do not limit the relationship between the specified number of samples expanded to the left and the specified number of samples expanded to the right, or the relationship between the specified number of samples expanded upwards and the specified number of samples expanded downwards. That is, the embodiments of this application do not limit the relationship between the range of the area expanded upwards by the encoder and the range expanded downwards by the encoder, or, that is, the embodiments of this application do not limit the relationship between the range of the area expanded to the left and the range expanded to the right by the encoder.

[0492] In an optional embodiment, the following example illustrates the situation: the number of specified sample points that the encoder expands upward is different from the number of specified sample points that the encoder expands downward, and the number of specified sample points that the encoder expands to the left is different from the number of specified sample points that the encoder expands to the right.

[0493] In one example, the number of samples the encoder adds upwards can be greater than the number of samples it adds downwards, and the number of samples it adds to the left can be greater than the number of samples it adds to the right. For example, the encoder adds 4 samples upwards, 2 samples downwards, 8 samples to the left, and 6 samples to the right.

[0494] In another example, the specified number of samples that the encoder expands upward can be greater than the specified number of samples that the encoder expands downward, and the specified number of samples that the encoder expands to the left can be less than the specified number of samples that the encoder expands to the right. For example, the encoder expands upward by 4 samples, expands downward by 2 samples, expands to the left by 6 samples, and expands to the right by 8 samples.

[0495] In another example, the specified number of samples the encoder adds upwards can be less than the specified number of samples the encoder adds downwards, and the specified number of samples the encoder adds to the left can be greater than the specified number of samples the encoder adds to the right. For example, the encoder adds 1 sample to the upward side, 5 samples to the downward side, 8 samples to the left, and 6 samples to the right.

[0496] In yet another example, the specified number of samples the encoder adds upwards can be less than the specified number of samples the encoder adds downwards, and the specified number of samples the encoder adds to the left can be less than the specified number of samples the encoder adds to the right. For example, the encoder adds 2 samples upwards, 4 samples downwards, 3 samples to the left, and 11 samples to the right.

[0497] Example B1

[0498] Corresponding to specific example 1 in example A1 above, specific example 1 in example B1

[0499] In conjunction with any of the above embodiments, this application also provides a decoding method, which may include the following steps:

[0500] Step 1: Obtain the bitstream of the image to be decoded.

[0501] In an optional embodiment, the image to be decoded includes a block to be decoded, which is the block data in the wavelet coefficients of the low-frequency subband obtained after wavelet transform during encoding of the image to be decoded. For example, the block to be decoded is the current block obtained in S102 shown in FIG20b, such as any one of blocks 1 to 9 in the image to be decoded shown in FIG21a(1).

[0502] When the image to be decoded is an original image (e.g., the input video or image shown in Figure 20b), the reference image is also an original image decoded before the image to be decoded. When the image to be decoded is a sub-image, the reference image is a sub-image decoded before the image to be decoded. The image to be decoded and the reference image correspond to each other.

[0503] The image to be decoded may include blocks to be decoded. A specific example 1-1 is illustrated in Figure 21a. Figure 21a(1) is a schematic diagram of a low-frequency sub-band of the current frame provided in an embodiment of this application. The low-frequency sub-band of the current frame may include nine data blocks to be decoded, arranged in the decoding order as block 1->block 2->block 3->block 4->block 5->block 6->block 7>block 8->block 9. For example, the current block on the decoding side may be block 5. Figure 21a(2) is a schematic diagram of a reconstructed low-frequency sub-band of a reference frame corresponding to the current frame provided in an embodiment of this application. The decoder has decoded nine data blocks within the reconstructed low-frequency sub-band of the reference frame, arranged in the decoding order as block 1'->block 2'->block 3'->block 4'->block 5'->block 6'->block 7'>block 8'->block 9'. Similarly, the encoding order and the decoding order can be the same, so the decoder can also decode each block in the current frame in the following order: block 1'-> block 2'-> block 3'-> block 4'-> block 5'-> block 6'-> block 7'> block 8'-> block 9'.

[0504] In related technologies, when performing inter-frame prediction, a reference image of the image to be decoded is used as a reference frame. Inter-frame prediction of the image to be decoded is performed by taking the pixel values ​​of pixels in the reference frame that are similar to those in the image to be decoded as the predicted values ​​of pixels in the image to be decoded, so as to obtain the prediction data for inter-frame prediction of the image to be decoded.

[0505] Step 2: Obtain the reconstructed low-frequency subband of the reference image corresponding to the image to be decoded.

[0506] In an optional embodiment, the reconstructed low-frequency subband is obtained based on the wavelet coefficients of the low-frequency subband obtained by wavelet transforming the reference image.

[0507] The reconstructed low-frequency subband is obtained based on the wavelet coefficients of the low-frequency subband obtained by wavelet transforming the reference image.

[0508] In some embodiments, the decoder can perform wavelet transform on the reference image to obtain the wavelet coefficients of the low-frequency sub-band; then, the reconstructed data of the wavelet coefficients of the low-frequency sub-band can be reconstructed based on the wavelet coefficients of the low-frequency sub-band, i.e., the low-frequency sub-band is reconstructed. The specific process can be seen in the process of obtaining the reconstructed low-frequency sub-band 1806 in Figure 18, or the process of obtaining the reconstructed low-frequency sub-band 1906 shown in Figure 19.

[0509] For a description of the acquired reconstructed low-frequency subband, please refer to the relevant description in Example A1 above. Optionally, the reconstructed low-frequency subband is an adjusted reconstructed low-frequency subband.

[0510] In this embodiment, any image (e.g., a reference image) can be transformed by wavelet transform to obtain a high-frequency sub-band and a low-frequency sub-band. Both the high-frequency and low-frequency sub-bands can include wavelet coefficients. Therefore, in subsequent embodiments, the wavelet coefficients in the low-frequency sub-band are described as low-frequency sub-bands.

[0511] In the following text, considering that the low-frequency subband may contain wavelet coefficients, for the sake of brevity, the process of encoding and decoding the wavelet coefficients of the low-frequency subband will be described using the operation flow of encoding and decoding the low-frequency subband. The physical meaning of encoding and decoding the low-frequency subband and encoding and decoding the wavelet coefficients of the low-frequency subband are equivalent.

[0512] Step 3: Based on the reconstructed low-frequency subband, perform inter-frame prediction on the block to be decoded to obtain the prediction data of the block to be decoded.

[0513] In some embodiments, the prediction data for the block to be decoded is data from a reference block obtained from reconstructing the low-frequency subband.

[0514] For example, during inter-frame prediction, the position of a reference block can be obtained within the reconstructed low-frequency subband based on the position of the block to be decoded and the MV of the block to be decoded, so that the data of the reference block in the reconstructed low-frequency subband can be used as the prediction block of the current block.

[0515] In an optional embodiment, the decoder obtains a reference block from the reconstructed low-frequency subband, and based on this reference block, the decoder obtains the prediction data for the block to be decoded. For example, the decoder determines the data of the reference block in the reconstructed low-frequency subband as the prediction data for the block to be decoded.

[0516] In this embodiment, considering that the size of the low-frequency subband output by the wavelet transform is only 1 / 4 of the image before the wavelet transform (e.g., the image to be decoded), obtaining the reference block based on the reconstructed low-frequency subband reduces the amount of data acquired compared to obtaining the reference block based on the reference frame, thereby improving inter-frame prediction efficiency and decoding efficiency. Furthermore, considering that the data distribution of the reconstructed low-frequency subband is more uniform than that of the reconstructed high-frequency subband output by the wavelet transform, determining the data from the reconstructed low-frequency subband as the prediction data for the block to be decoded results in more accurate prediction data, improving the accuracy of inter-frame prediction and thus enhancing decoding accuracy.

[0517] This method can be implemented in conjunction with the wavelet transform-based decoder described above, as shown in the flowcharts in Figures 7, 11, 18, and 19. In one example, step 3 can be implemented using prediction unit 1814 as shown in Figure 18, or prediction unit 1914 as shown in Figure 19. The specific principles can be found in the descriptions of prediction units 1814 and 1914 above, and will not be repeated here.

[0518] Step 4: Based on the predicted data, decode the bitstream to obtain the reconstructed data of the block to be decoded.

[0519] In some embodiments, the specific decoding process can refer to the process of decoding the current block (e.g., macroblock 1204) based on the prediction block, as shown in Figures 7, 11, 18, and 19 above, and will not be elaborated here.

[0520] For example, the low-frequency sub-band of the current frame shown in Figure 21a(1) can be the low-frequency sub-band of the image to be decoded. The low-frequency sub-band of the image to be decoded can include multiple blocks of data, and the bitstream can be decoded according to the decoding order of the block data to obtain the reconstructed data of each block.

[0521] The reconstructed low-frequency subband of the reference frame shown in Figures 21a(2) and 21b can be the reconstructed low-frequency subband obtained in step 3.

[0522] The specific embodiments of Figures 21a and 21b can also be applied to the decoding method of this application, such as steps 2 and 3 above, as the principle is the same, and will not be repeated here.

[0523] Compared to acquiring a reference frame for inter-frame prediction of the current block within the current frame, this embodiment acquires the reconstructed low-frequency subband data (i.e., reconstructed low-frequency subband) of the reference frame to perform inter-frame prediction of the current block within the current frame. The size of the low-frequency subband of the reference frame is one-quarter of the reference frame's size, thus significantly reducing the amount of reference data acquired during inter-frame prediction, and consequently reducing the bit width occupied by the reference data. Since it is not necessary to read the entire image data of the reference frame, but only its reconstructed low-frequency subband, memory usage and bandwidth consumption can be reduced. Furthermore, performing inter-frame prediction of the current block based on the reconstructed low-frequency subband of the reference frame to obtain the predicted value of the current block for decoding the current frame improves the efficiency of inter-frame prediction compared to performing inter-frame prediction based on the reference frame, thereby improving decoding efficiency.

[0524] After wavelet transform, the low-frequency subbands (LL subbands) of adjacent frames still have a large amount of redundancy in the time domain, which can be further compressed through inter-frame prediction. Compared to reading the reconstructed data of the original-size reference image as the reference frame, the embodiments of this application perform inter-frame prediction in the dimension of the LL subband. Only the reconstructed low-frequency subband of the reference image needs to be read as the reference data of the current frame. The size of the reconstructed low-frequency subband is 1 / 4 of the size of the reference image, which reduces the read-in bandwidth consumption.

[0525] Based on any of the above wavelet transform-based embodiments (e.g., embodiments of any of the decoding methods in Figures 6 to 19) and any of the decoding methods involving inter-frame prediction in Figures 1 to 5, the decoding methods of this application embodiments will be described below with reference to different examples.

[0526] Corresponding to Specific Example 2 in Example A1, this application provides Specific Example 2 in Example B1.

[0527] Figure 23 illustrates a flowchart of a decoding method according to this application. As shown in Figure 23, the method flow may include, but is not limited to, the following steps:

[0528] S401. Decode the bitstream of the current frame.

[0529] In an optional embodiment, the decoder decodes the bitstream to obtain the position of the current block (optionally obtained from the bitstream, or determined according to the decoding order so as not to be obtained from the bitstream), the residual of the current block, and the MVD of the current block.

[0530] S402. Obtain the reconstructed low-frequency subband of the reference frame corresponding to the current frame.

[0531] In an optional embodiment, the reference frame is a frame that has already been decoded. After the decoder decodes the reference frame, it can reconstruct the data of the decoded reference frame to obtain the reconstructed value of the reference frame. The decoder can read the reconstructed value of the reference frame.

[0532] The specific implementation principle of S402 is the same as that of step 2 in specific example 1 in example B1 above, and will not be repeated here.

[0533] S403. Based on the reconstructed low-frequency subband, perform inter-frame prediction on the current block to obtain the predicted value of the current block.

[0534] This application does not limit the specific decoding process of the predicted value of the current block obtained by inter-frame prediction. It can be the decoding implementation process of the inter-frame prediction result of any of the above embodiments. This application also does not limit the syntax elements in the bitstream. Specifically, it can be combined with the decoding method or decoder involving wavelet transform in any of the above embodiments for decoding.

[0535] In conjunction with any of the above-described decoding-side embodiments, FIG24 exemplarily illustrates a flowchart of inter-frame prediction according to this application. For example, the specific implementation process of S403 in FIG23 can be implemented through the flowchart in FIG24.

[0536] As shown in Figure 24, the method flow may include, but is not limited to, the following steps:

[0537] S431. Based on the MV of the decoded block in the low-frequency subband of the current frame and the MVD of the current block, obtain the MV of the current block.

[0538] Referring to Figure 23, the MVD of the current block is obtained from the bitstream decoding. In addition, the decoder can also obtain the MV of the decoded blocks in the low-frequency subband of the current frame (e.g., decoded blocks adjacent to the current block). Then, based on the MV of the decoded blocks and the MVD of the current block, the decoder can obtain the MV of the current block.

[0539] S432. Based on the MV of the current block and the reconstructed low-frequency subband of the reference frame (e.g., obtained from S402 as shown in Figure 23), obtain the prediction value of the current block.

[0540] In an optional embodiment, the predicted value of the current block is the data of the reference block obtained from the reconstructed low-frequency subband of the reference frame.

[0541] In a specific embodiment, the decoder can obtain a reference block from the reconstructed low-frequency subband of the reference frame based on the MV of the current block, and the decoder uses the data of the obtained reference block as the prediction value of the current block.

[0542] For example, the decoder can obtain the position of the reference block based on the MV of the current block and the position of the current block; then, based on the reconstructed low-frequency subband of the reference frame and the position of the reference block, it can use the data corresponding to the position of the reference block (e.g., obtained from the reconstructed low-frequency subband) as the prediction value of the current block.

[0543] In some embodiments, if the current block on the decoding side undergoes wavelet transform and adjustment operations during the above encoding, then in order to ensure accurate decoding, in this embodiment, adjustment operations can be performed on the reconstructed low-frequency subband of the reference frame as shown in FIG24. This adjustment operation is the same as the adjustment operation performed on the reconstructed low-frequency subband of the reference frame on the encoding side.

[0544] Thus, in step S432 as shown in Figure 24, the position of the reference block can be determined within the reconstructed low-frequency subband after adjustment based on the MV of the current block and the position of the current block. The data of the reference block located at that position within the reconstructed low-frequency subband after adjustment (also referred to as the reconstructed value of the reference block) is then used as the predicted value of the current block.

[0545] In some embodiments, where the adjustment operation includes precision adjustment, the decoder may perform precision adjustment (e.g., left shift by 2 bits) on the reconstructed low-frequency subband of the reference frame before S432, and then perform step S432 on the precision-adjusted reconstructed low-frequency subband.

[0546] In some embodiments, where the adjustment operation includes the aforementioned range adjustment, the decoder may perform range adjustment (e.g., increase the offset) on the reconstructed low-frequency subband of the reference frame before S432, and then perform step S432 on the range-adjusted reconstructed low-frequency subband. Alternatively, after S432 as shown in FIG24, the prediction value of the current block (equivalent to not having undergone range adjustment) may be obtained, and then the prediction value of the current block may be range-adjusted so that the range-adjusted prediction value of the current block and the residual of the current block decoded from the bitstream have the same data dimension. Since the prediction value is data obtained from the reconstructed low-frequency subband of the reference frame, performing range adjustment on the prediction value is equivalent to performing range adjustment on the data of the reconstructed low-frequency subband.

[0547] After obtaining the predicted value of the current block through inter-frame prediction, the range of the predicted value of the current block is then adjusted. Compared with adjusting the range of the reconstructed low-frequency subband of the reference frame before inter-frame prediction and then using it for inter-frame prediction, the memory usage of inter-frame prediction can be reduced and the prediction efficiency can be improved.

[0548] In some embodiments, the predicted value of the current block is data of a reference block obtained from a first region of the reconstructed low-frequency subband of a reference frame. This first region is obtained by expanding a specified number of samples around the corresponding position of the current block in the reconstructed low-frequency subband (hereinafter referred to as the "reference position").

[0549] In an optional embodiment, the decoder expands the corresponding position of the current block in the reconstructed low-frequency subband by a specified number of samples to obtain a first region. This first region is the region determined in the reconstructed low-frequency subband. Then, based on the position and MV of the current block, the decoder can determine a reference block similar to the current block within the first region of the reconstructed low-frequency subband of the reference frame, and use the data of the reference block in the first region as the prediction value of the current block.

[0550] The reference block data is the data of an 8x8 block within the reconstructed low-frequency subband (which may be the data after padding the reconstructed low-frequency subband according to the aforementioned "expansion"). The decoder can locate the 8x8 reference block within the first region of the reconstructed low-frequency subband, based on the current block's MV and its position.

[0551] For example, the decoder shown in FIG18 can implement the decoding method of the present application embodiment. Specifically, the prediction unit 1814 shown in FIG18 can perform inter-frame prediction on the block to be decoded based on the reconstructed low-frequency subband 1806 of the reference image to obtain the prediction block 1805.

[0552] For example, the decoder shown in FIG19 can implement the decoding method of the present application embodiment. Specifically, the prediction unit 1914 shown in FIG19 can perform inter-frame prediction on the block to be decoded based on the reconstructed low-frequency subband 1906 of the reference image to obtain the prediction block 1905.

[0553] The specific process of inter-frame prediction can be implemented by prediction unit 1814 as shown in Figure 18 or prediction unit 1914 as shown in Figure 19, following the inter-frame prediction process. This application does not impose any restrictions.

[0554] For example, prediction unit 1814 (or prediction unit 1914) may include an inter-frame prediction unit (not shown), which may include a motion estimation (ME) unit (not shown in Figures 18 and 19) and a motion compensation (MC) unit (not shown in Figures 18 and 19). The functions of the motion estimation unit and the motion compensation unit can be referred to the description of the above embodiments.

[0555] In a specific embodiment, under the same coordinate system, the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame is the same as the position of the current block in the low-frequency subband of the current frame.

[0556] In an optional embodiment, the decoder may determine the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame based on the position of the current block in the low-frequency subband of the current frame (e.g., the low-frequency subband obtained by wavelet transform of the current frame).

[0557] Referring to the specific example 1-1 above, and referring to Figure 21a, the size of each block is 8x8. Referring to Figure 21a(1), the block to be decoded (i.e., the current block) in the low-frequency subband of the current frame is block 5. For example, the corresponding position of block 5 in the reconstructed low-frequency subband of the reference frame shown in Figure 21a(2) (i.e., the above-mentioned reference position) is the position of block 5' in the reconstructed low-frequency subband of the reference frame.

[0558] In a specific embodiment, the decoder can expand a specified number of sample points around the position of block 5' as shown in Figure 21a(2) to obtain the first region 10.

[0559] As shown in Figure 21a(2), the surrounding area may include at least one side of the horizontal direction starting from the position of block 5', and / or at least one side of the vertical direction starting from the position of block 5'. Thus, the decoder expands to at least one side of the horizontal left and right sides of the position of block 5' (e.g., a plurality of arrows on the left and right sides of block 5', expanding the left and right sides in the direction of the arrows) and to at least one side of the vertical top and bottom sides of the position of block 5' (e.g., a plurality of arrows on the top and bottom sides of block 5', expanding the top and bottom sides in the direction of the arrows) to obtain the first region 10 shown in the dashed box.

[0560] Of course, the number of sample points expanded in the four directions of up, down, left, and right can be the same or different, and there is no restriction here.

[0561] Furthermore, the shape of the first region is not limited to a rectangle; it can also be an irregular shape.

[0562] Then, as shown in Figure 21a(2), the decoder can determine the position of an 8x8 reference block within the first region 10 that has a pixel value close to that of block 5 (an example of the current block) as shown in Figure 21a(1), based on the MV of the current block (e.g., block 5) and the position of the current block. The block data of the 8x8 reference block within the reconstructed low-frequency subband of the reference frame is then used as the prediction block for inter-frame prediction of block 5 as shown in Figure 21a(1). The position of the current block (e.g., block 5) is the position of block 5' in the first region 10 of the reconstructed low-frequency subband as shown in Figure 21a(1).

[0563] In this way, when the decoder performs inter-frame prediction on the current block within the current frame based on the reconstructed low-frequency subband of the reference frame, it can obtain the reference block data in the first region of the reconstructed low-frequency subband, which can then be used as the prediction data for the inter-frame prediction of the current block. This ensures that the reference block is not far from the current block, but rather within the controllable first region, thereby improving the efficiency of obtaining the prediction block, thus improving the prediction efficiency of inter-frame prediction and reducing the amount of decoding data to improve decoding efficiency. Furthermore, the first region is obtained by expanding the area around the corresponding position of the current block in the reconstructed low-frequency subband of the reference frame, ensuring that this region is still near the corresponding position of the current block. This means that the reference block obtained by inter-frame prediction is also quite similar to the current block, ensuring the accuracy of inter-frame prediction, thereby reducing the amount of decoding data and improving decoding efficiency. In this embodiment, performing inter-frame prediction within the first region to obtain the prediction data of the current block reduces decoding complexity compared to determining the prediction data of the current block within the entire reference frame.

[0564] In some embodiments, the first region is a region comprising the reconstructed low-frequency subband and expanded around the reconstructed low-frequency subband by a specified number of samples. For example, as shown in FIG21c, the reconstructed low-frequency subband includes multiple blocks, such as block 1', block 2', etc. The reconstructed low-frequency subband is expanded in the horizontal and / or vertical directions to obtain the first region 13. In this embodiment, the relevant description of the first region, the expansion method, and the number of expansion samples can be referred to the relevant description in specific example 2 of example A1 above, and will not be repeated here.

[0565] In an optional embodiment, the decoder may obtain data of the reference block based on the reconstructed low-frequency subband from the reference frame as the prediction value of the current block.

[0566] In an optional embodiment, the predicted value of the current block is the reconstructed value of the reference block obtained from a first region of the reference frame. The first region is a region obtained by expanding the corresponding position of the current block in the reference frame by a specified number of samples. The corresponding position of the current block in the reference frame is also called the reference position.

[0567] Returning to Figure 23, the method also includes S404.

[0568] S404. Based on the residual of the current block and the predicted value of the current block, obtain the reconstruction value of the current block.

[0569] In an optional embodiment, the decoder determines the reconstructed value of the current block by summing the residual of the current block with the predicted value of the current block.

[0570] In an optional embodiment, the first region is a region obtained by expanding the same number of first reference points to both sides in the horizontal direction based on the above-mentioned "reference position".

[0571] In a specific embodiment, the decoder expands the first number of identical points horizontally to the left and right from the "reference position" to obtain the first region. For example, the first number of identical points is 7. In other embodiments, the first number of identical points can be any integer greater than 0, and can be flexibly set according to needs without limitation.

[0572] The embodiments of this application do not limit the order in which the decoder expands to both sides in the horizontal direction.

[0573] In this embodiment of the application, since the first region is the region used to determine the position of the reference block for inter-frame prediction of the current block, the first region can be set as the region obtained by expanding the same number of samples to the left and right sides in the horizontal direction with the corresponding position of the current block in the reconstructed low-frequency sub-band as the center.

[0574] In an optional embodiment, the first region is a region obtained by expanding the same number of second sample points to both sides in the vertical direction based on the above-mentioned "reference position".

[0575] In a specific embodiment, the decoder expands the number of second sample points vertically upwards and downwards from the "reference position" to obtain the first region. For example, the number of second sample points is 3. In other embodiments, the number of second sample points can be any integer greater than 0, and can be flexibly set according to needs without limitation.

[0576] This application does not limit the order in which the decoder expands to both sides in the vertical direction. In a specific embodiment, the decoder expands the second number of sample points downwards in the vertical direction from the "reference position," and then expands the second number of sample points upwards in the vertical direction to obtain the first region. For example, the second number of sample points is 3.

[0577] In some embodiments, in order to reduce the bitstream and improve encoding and decoding efficiency, the specific values ​​of the first and second sample points can be preset on the decoder (and encoding side) side without being transmitted from the bitstream.

[0578] In an optional embodiment, the first region is a region obtained by expanding the number of sample points upward in the vertical direction by D1 and the number of sample points downward by D2 based on the "reference position".

[0579] In a specific embodiment, the decoder expands the number of sample points D1 vertically upward from the "reference position" and then expands the number of sample points D2 vertically downward to obtain the first region.

[0580] In a specific embodiment, the decoder expands the number of sample points D2 downwards in the vertical direction from the "reference position" and then expands the number of sample points D1 upwards in the vertical direction to obtain the first region.

[0581] It should be noted that the embodiments of this application do not impose restrictions on the specific values ​​of the number of sample points D1 and D2, as long as both the number of sample points D1 and D2 are positive integers greater than or equal to 0.

[0582] It should be noted that the embodiments of this application do not limit the size relationship between the number of sample points D1 and the number of sample points D2. That is, the embodiments of this application do not limit the size relationship between the range of the area expanded upward by the decoder and the range expanded downward by the decoder.

[0583] In one example, the number of samples D1 can be greater than the number of samples D2. For example, the number of samples D1 can be 4 and the number of samples D2 can be 2. In another example, the number of samples D1 can also be less than the number of samples D2. For example, the number of samples D1 can be 1 and the number of samples D2 can be 5.

[0584] In an optional embodiment, the decoder first expands the corresponding position of the current block by a specified number of samples horizontally, and then expands it by a specified number of samples vertically to obtain a first region. Alternatively, the decoder first expands the corresponding position of the current block by a specified number of samples vertically, and then expands it by a specified number of samples horizontally to obtain a first region.

[0585] It should be noted that the embodiments of this application do not limit whether the specified number of samples expanded upward by the decoder is the same as the specified number of samples expanded downward by the decoder; and the embodiments of this application do not limit whether the specified number of samples expanded to the left by the decoder is the same as the specified number of samples expanded to the right by the decoder. That is, the embodiments of this application do not limit whether the range of the decoder expanded upward is equal to the range of the decoder expanded downward; and the embodiments of this application do not limit whether the range of the decoder expanded to the left is equal to the range of the decoder expanded to the right.

[0586] In some embodiments, the specified number of samples that the decoder expands upwards can be the same as the specified number of samples that the decoder expands downwards, and the specified number of samples that the decoder expands to the left can be the same as the specified number of samples that the decoder expands to the right. For example, the decoder expands upwards and downwards by 3 samples each, and expands to the left and right by 7 samples each.

[0587] It should be noted that the embodiments of this application do not limit the relationship between the specified number of samples expanded in the horizontal direction and the specified number of samples expanded in the vertical direction. That is, the embodiments of this application do not limit the relationship between the range of the area expanded by the decoder in the horizontal direction and the range of the area expanded by the decoder in the vertical direction.

[0588] In some embodiments, the first sample number is greater than the second sample number. For example, the first region determined by the decoder is the region located in the reconstructed low-frequency subband obtained by expanding the number of samples in the horizontal direction with the "reference position" as the center, which is greater than the number of samples expanded in the vertical direction.

[0589] The process of determining the first region and the padding process of the first region can be referred to the description of the embodiments of Figures 21a, 21b and 21c on the encoding side. The principle is the same, so it will not be repeated here.

[0590] Thus, the number of samples expanded horizontally is greater than the number of samples expanded vertically. In this embodiment, considering that the decoder reads data in the reconstructed low-frequency subband of the reference frame line by line, if the number of samples is greater than the number of samples, the decoder will read more lines of data in the first region of the low-frequency subband during line-by-line reading, resulting in excessive memory consumption. Therefore, in this embodiment, the specified number of samples expanded horizontally in the first region is greater than the specified number of samples expanded vertically by the decoder, which reduces the number of lines of data read by the decoder in the first region, thereby reducing memory consumption during inter-frame prediction.

[0591] In the embodiment of Figure 21a(1), the size of the block is 8x8, and it is expanded by 7 samples on both sides in the horizontal direction and by 3 samples on both sides in the vertical direction to obtain the first region. The first region obtained in this way can be the region with the highest probability of the reference block appearing in the inter-frame prediction of the current block. Then, inter-frame prediction is performed in the first region to obtain the prediction block of the current block, which can improve the prediction efficiency and prediction accuracy of the current block, thereby improving the decoding speed and decoding efficiency.

[0592] In the embodiment of FIG21b(3), when the reconstructed low-frequency subband of the reference frame cannot cover the first region, reference data in the first region (i.e., data in the reconstructed low-frequency subband located in the first region, such as data in region 12 shown in FIG21b(3)) can be obtained by filling data around the "reference position" based on the data of the reconstructed low-frequency subband. Obviously, the reference data in the first region 12 shown in FIG21b(3) is also obtained based on the reconstructed low-frequency subband. So when the reconstructed low-frequency subband of the reference frame cannot cover the first region, since the additional data filled in the first region is based on the data of the reconstructed low-frequency subband, the first region in each embodiment of this application can also be described as the first region of the reconstructed low-frequency subband (or described as the first region within the reconstructed low-frequency subband). Similarly, when the reconstructed low-frequency subband of the reference frame cannot cover the first region, since the additional data filled in the first region is based on the data of the reconstructed low-frequency subband, the prediction data of the current block obtained in the first region can also be described as the data of the reference block obtained from the reconstructed low-frequency subband.

[0593] In some embodiments, when the codec hardware performance is strong, the number of samples added horizontally can be less than the number of samples added vertically. For example, the number of samples added horizontally is 10, and the number of samples added vertically is 12. That is, the number of samples added to the left is 5, the number of samples added to the right is 5, the number of samples added upwards is 6, and the number of samples added downwards is 6.

[0594] In another embodiment, the specified number of samples the decoder adds horizontally can be equal to the specified number of samples the decoder adds vertically. For example, the number of samples added horizontally is 10, and the number of samples added vertically is 10. That is, the number of samples added to the left, the number of samples added to the right, the number of samples added upwards, and the number of samples added downwards are all 5.

[0595] In other embodiments, the specified number of samples that the decoder expands upwards can be the same as the specified number of samples that the decoder expands downwards, and the specified number of samples that the decoder expands to the left can be different from the specified number of samples that the decoder expands to the right. For example, the decoder expands upwards and downwards by 3 samples each, expands to the left by 8 samples, and expands to the right by 6 samples.

[0596] In some other embodiments, the specified number of samples that the decoder expands upwards may be different from the specified number of samples that the decoder expands downwards, and the specified number of samples that the decoder expands to the left may be the same as the specified number of samples that the decoder expands to the right. For example, the decoder expands upwards by 4 samples, expands downwards by 2 samples, and expands to the left and right by 7 samples each.

[0597] In some other embodiments, the specified number of samples that the decoder expands upwards may be different from the specified number of samples that the decoder expands downwards, and the specified number of samples that the decoder expands to the left may be different from the specified number of samples that the decoder expands to the right. For example, the decoder expands upwards by 4 samples, expands downwards by 2 samples, expands to the left by 8 samples, and expands to the right by 6 samples.

[0598] It should be understood that when the number of specified samples expanded upwards by the decoder differs from the number of specified samples expanded downwards, and / or when the number of specified samples expanded to the left differs from the number of specified samples expanded to the right, the embodiments of this application do not limit the relationship between the number of specified samples expanded to the left and the number of specified samples expanded to the right, or the relationship between the number of specified samples expanded upwards and the number of specified samples expanded downwards. That is, the embodiments of this application do not limit the relationship between the range of the region expanded upwards by the decoder and the range expanded downwards by the decoder, or, that is, the embodiments of this application do not limit the relationship between the range of the region expanded to the left and the range expanded to the right by the decoder.

[0599] In an optional embodiment, the following example illustrates the situation: the number of specified samples that the decoder expands upwards is different from the number of specified samples that the decoder expands downwards, and the number of specified samples that the decoder expands to the left is different from the number of specified samples that the decoder expands to the right.

[0600] In one example, the number of samples the decoder adds upwards can be greater than the number of samples it adds downwards, and the number of samples it adds to the left can be greater than the number of samples it adds to the right. For example, the decoder adds 4 samples upwards, 2 samples downwards, 8 samples to the left, and 6 samples to the right.

[0601] In another example, the number of samples the decoder expands upward can be greater than the number of samples the decoder expands downward, and the number of samples the decoder expands to the left can be less than the number of samples the decoder expands to the right. For example, the decoder expands upward by 4 samples, expands downward by 2 samples, expands to the left by 6 samples, and expands to the right by 8 samples.

[0602] In another example, the number of samples the decoder adds upwards can be less than the number of samples it adds downwards, and the number of samples it adds to the left can be greater than the number of samples it adds to the right. For example, the decoder adds 1 sample upwards, 5 samples downwards, 8 samples to the left, and 6 samples to the right.

[0603] In yet another example, the number of samples the decoder adds upwards can be less than the number of samples it adds downwards, and the number of samples it adds to the left can be less than the number of samples it adds to the right. For example, the decoder adds 2 samples upwards, 4 samples downwards, 3 samples to the left, and 11 samples to the right.

[0604] Based on the same concept as the above method, as shown in FIG25a, this application embodiment also provides a decoding device 1000, which includes: an acquisition module 1001, used to acquire the bitstream of an image to be decoded, the image to be decoded including a block to be decoded; the acquisition module 1001 is further used to acquire the reconstructed low-frequency subband of a reference image corresponding to the image to be decoded, the reconstructed low-frequency subband being obtained based on the wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image; a prediction module 1002, used to perform inter-frame prediction on the block to be decoded based on the reconstructed low-frequency subband to obtain prediction data of the block to be decoded; and a decoding module 1003, used to decode the bitstream based on the prediction data to obtain reconstructed data of the block to be decoded.

[0605] Based on the same concept as the above method, as shown in Figure 25b, this application embodiment also provides an encoding device 2000, which includes: an acquisition module 2001, used to acquire the reconstructed low-frequency subband of a reference image corresponding to the image to be encoded, wherein the reconstructed low-frequency subband is obtained based on the wavelet coefficients of the low-frequency subband obtained by wavelet transform of the reference image, and the image to be encoded includes a block to be encoded; a prediction module 2002, used to perform inter-frame prediction on the block to be encoded based on the reconstructed low-frequency subband to obtain prediction data of the block to be encoded; and an encoding module 2003, used to encode the block to be encoded based on the prediction data to obtain a bitstream.

[0606] The prediction module 1002 of the aforementioned decoding device can be applied to the inter-frame prediction process at the decoding end. Specifically, at the decoding end, the prediction module 1002 can be applied to the inter-frame prediction unit or prediction unit of the aforementioned decoder.

[0607] The prediction module 2002 of the aforementioned encoding device can be applied to the inter-frame prediction process at the encoding end. Specifically, at the encoding end, the prediction module 2002 can be applied to the inter-frame prediction unit or prediction unit of the aforementioned encoder.

[0608] The specific implementation process of the encoding device and the decoding device can be referred to the relevant descriptions of the encoding method and decoding method embodiments, and will not be repeated here for the sake of brevity.

[0609] The steps of the methods or algorithms described in conjunction with the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium well known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0610] Those skilled in the art will recognize that the functions described in the embodiments of this application in one or more of the above examples can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of a computer program from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0611] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of this application.

Claims

1. A decoding method, characterized in that, The method includes: Obtain the bitstream of the image to be decoded, wherein the image to be decoded includes blocks to be decoded; Based on the reconstructed low-frequency subband of the reference image corresponding to the image to be decoded, inter-frame prediction is performed on the block to be decoded to obtain the prediction data of the block to be decoded. Based on the predicted data, the bitstream is decoded to obtain the reconstructed data of the block to be decoded.

2. The method according to claim 1, characterized in that, The block to be decoded is the block data of the low-frequency subband of the image to be decoded.

3. The method according to claim 1, characterized in that, The prediction data is data from the reference block obtained from the reconstructed low-frequency subband.

4. The method according to any one of claims 1 to 3, characterized in that, The predicted data is data from a reference block obtained from a first region of the reconstructed low-frequency subband, the first region being a region that includes the reconstructed low-frequency subband and is expanded around the reconstructed low-frequency subband by a specified number of samples.

5. The method according to claim 4, characterized in that, The first region is the region obtained by including the reconstructed low-frequency subband and expanding the same number of first points to both sides in the horizontal direction.

6. The method according to claim 4 or 5, characterized in that, The first region is the region obtained by including the reconstructed low-frequency subband and expanding the same number of second sample points to both sides in the vertical direction.

7. The method according to claim 6, characterized in that, The number of the first sample points is greater than the number of the second sample points.

8. The method according to claim 7, characterized in that, The number of the first sample points is not less than 7, and the number of the second sample points is not less than 3.

9. An encoding method, characterized in that, The method includes: Obtain the reconstructed low-frequency subband of the reference image corresponding to the image to be encoded; the image to be encoded includes the block to be encoded. Based on the reconstructed low-frequency subband, inter-frame prediction is performed on the block to be coded to obtain the prediction data of the block to be coded. Based on the predicted data, the block to be encoded is encoded to obtain a bitstream.

10. The method according to claim 9, characterized in that, The prediction data is data from the reference block obtained from the reconstructed low-frequency subband.

11. The method according to claim 9 or 10, characterized in that, The predicted data is data from a reference block obtained from a first region of the reconstructed low-frequency subband, the first region being a region that includes the reconstructed low-frequency subband and is expanded around the reconstructed low-frequency subband by a specified number of samples.

12. The method according to claim 11, characterized in that, The first region is the region obtained by including the reconstructed low-frequency subband and expanding the same number of first points to both sides in the horizontal direction.

13. The method according to claim 11 or 12, characterized in that, The first region includes the region obtained by reconstructing the low-frequency subband and expanding it to both sides in the vertical direction with the same number of second sample points.

14. The method according to claim 13, characterized in that, The number of the first sample points is greater than the number of the second sample points.

15. The method according to claim 14, characterized in that, The number of the first sample points is not less than 7, and the number of the second sample points is not less than 3.

16. The method according to any one of claims 9 to 15, characterized in that, The block to be encoded is the low-frequency subband block data obtained by wavelet transforming the image to be encoded.

17. A bitstream, characterized in that, The bitstream is generated according to the method described in any one of claims 9 to 16 above.

18. An encoder, characterized in that, include: A memory and a processor, wherein the memory is coupled to the processor; The memory stores program instructions that, when executed by the processor, cause the processor to perform the steps of the method as described in any one of claims 9 to 16.

19. A decoder, characterized in that, include: A memory and a processor, wherein the memory is coupled to the processor; The memory stores program instructions that, when executed by the processor, cause the processor to perform the steps of the method as described in any one of claims 1 to 8.

20. An encoder, characterized in that, include: A processing circuit that implements the steps of the method as claimed in any one of claims 9 to 16.

21. A decoder, characterized in that, include: A processing circuit that implements the steps of the method as claimed in any one of claims 1 to 8.

22. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed on a computer or processor, causes the computer or processor to perform the method as described in any one of claims 1 to 8, or causes the computer or processor to perform the method as described in any one of claims 9 to 16.

23. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed by a computer or processor, cause the steps of the method as described in any one of claims 1 to 8 to be performed, or cause the steps of the method as described in any one of claims 9 to 16 to be performed.

24. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a bitstream generated according to the method described in any one of claims 9 to 16.

Images