Image decoding and encoding method, apparatus, device, and storage medium

Abstract

The present invention discloses an image decoding and encoding method, apparatus, device, and storage medium, which includes decoding an image bitstream, determining feature reconstruction values ​​corresponding to current image features obtained by decoding, performing feature enhancement on the feature reconstruction values ​​to obtain enhanced feature values, and performing synthesis transformation on the enhanced feature values ​​to obtain reconstructed image blocks.

Description

[Technical Field]

[0001] This application claims priority from a Chinese patent application bearing application number 202310055970.0, filed on January 13, 2023, the entire contents of which are incorporated herein by reference.

[0002] The present invention relates to the field of image processing technology, and in particular to an image decoding and encoding method, apparatus, device, and storage medium. [Background technology]

[0003] Currently, deep learning and neural networks are making great strides in the field of video image compression, and image coding techniques based on deep learning have significantly surpassed conventional coding standards in terms of coding performance. However, how to improve the image quality of images reconstructed after compression using image coding techniques based on deep learning remains a major challenge.

[0004] The above content is only an aid for understanding the technical solution of the present invention, and is not intended to be an admission that the above content is prior art. Summary of the Invention [Means for solving the problem]

[0005] The main object of the present invention is to provide an image decoding and encoding method, apparatus, device, and storage medium that improves the image quality of an image reconstructed after being compressed using deep learning-based image encoding techniques.

[0006] In order to achieve the above object, the present invention provides: decoding the image bitstream and determining feature reconstruction values ​​corresponding to the current image features obtained by the decoding; performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value; and performing a synthesis transform on the enhancement feature values ​​to obtain a reconstructed image block. An image decoding method is provided.

[0007] In order to achieve the above object, the present invention further provides: a bitstream decoding module for decoding the image bitstream and determining a feature reconstruction value corresponding to the current image feature obtained by the decoding; a feature enhancement module for performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value; an image reconstruction module for performing a synthesis transform on the enhancement feature values ​​to obtain a reconstructed image block. An image decoding device is provided.

[0008] In order to achieve the above object, the present invention further provides: performing feature extraction on the image block to be encoded and setting the extracted features as current image features; performing prediction based on feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values; determining coding residual coefficients corresponding to the current image features based on the predicted feature values; writing the coding residual coefficients into an image bitstream corresponding to the image block to be coded. An image coding method is provided.

[0009] In order to achieve the above object, the present invention further provides: a feature extraction module for performing feature extraction on the image block to be encoded and setting the extracted features as current image features; a feature prediction module for performing prediction based on feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values; a residual calculation module for determining coding residual coefficients corresponding to the current image feature based on the predicted feature values; a parameter writing module for writing the coding residual coefficients into an image bitstream corresponding to the image block to be coded. An image encoding device is provided.

[0010] In order to achieve the above object, the present invention further provides: a decoding device including a processor, a memory, and a decoding program stored in the memory and executable by the processor, the decoding device performing the image decoding method described above when the decoding program is executed by the processor; A decryption device is provided.

[0011] In order to achieve the above object, the present invention further provides: An encoding device including a processor, a memory, and a decoding program and / or an encoding program stored in the memory and executable by the processor, wherein the decoding program is executed by the processor to implement the image decoding method, and the encoding program is executed by the processor to implement the image encoding method. An encoding device is provided.

[0012] In order to achieve the above object, the present invention further provides: A computer-readable storage medium storing an image decoding program and / or an image encoding program, wherein the image decoding program is executed to implement the image decoding method, and the image encoding program is executed to implement the image encoding method. A computer-readable storage medium is provided.

[0013] The present invention decodes an image bitstream, determines feature reconstruction values ​​corresponding to the current image features obtained by decoding, performs feature enhancement on the feature reconstruction values ​​to obtain enhanced feature values, and performs synthesis transformation on the enhanced feature values ​​to obtain reconstructed image blocks. Before performing image reconstruction, feature enhancement is performed on the feature reconstruction values, and then image reconstruction is performed based on the enhanced feature enhancement values, thereby reducing distortion of image features in processes such as quantization and improving the quality of the reconstructed image. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a schematic structural diagram of an electronic device in a hardware execution environment according to a technical solution of an embodiment of the present invention; [Figure 2] 1 is a schematic flowchart of a first embodiment of an image decoding method of the present invention. [Figure 3] FIG. 2 is a schematic diagram of a matrix structure according to an embodiment of the present invention. [Figure 4] 1 is a schematic flowchart of image encoding and decoding according to an embodiment of the present invention. [Figure 5] 1 is a schematic flowchart of image encoding and decoding according to an embodiment of the present invention. [Figure 6] 10 is a schematic flowchart of a second embodiment of the image decoding method of the present invention. [Figure 7] 10 is a schematic flowchart of a third embodiment of the image decoding method of the present invention. [Figure 8] 10 is a schematic flowchart of a fourth embodiment of the image decoding method of the present invention. [Figure 9] FIG. 2 is a schematic diagram of a feature reconstruction order for one embodiment of the present invention. [Figure 10] 1 is a schematic flowchart of a first embodiment of an image coding method of the present invention. [Figure 11] 10 is a schematic flowchart of a second embodiment of the image coding method of the present invention. [Figure 12] FIG. 1 is a structural block diagram of a first embodiment of an image decoding device according to the present invention. [Figure 13] 1 is a structural block diagram of a first embodiment of an image encoding device according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0015] The realization of the objects, functional features and advantages of the present invention will be further explained with reference to the drawings in combination with the embodiments.

[0016] It should be understood that the specific examples described herein are used only to illustrate the present invention and are not intended to limit the present invention.

[0017] Please refer to FIG. 1, which is a schematic structural diagram of a decoding device or an encoding device in a hardware execution environment according to a technical solution of an embodiment of the present invention.

[0018] As shown in FIG. 1 , the electronic device may include a processor 1001 such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to realize communication between these components. The user interface 1003 may include input units such as a display and a keyboard, and may further include a standard wired interface or a wireless interface. The network interface 1004 may include a standard wired interface or a wireless interface (e.g., a wireless fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM) such as a magnetic disk memory. The memory 1005 may be a storage device independent of the processor 1001.

[0019] As will be appreciated by those skilled in the art, the structure shown in FIG. 1 is not limiting of electronic devices that may include more or fewer components, some combinations of components, or different arrangements of components.

[0020] As shown in FIG. 1, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a decoding and / or encoding program.

[0021] In the electronic device shown in Figure 1, the network interface 1004 is mainly used for data communication with a network server, and the user interface 1003 is mainly used for data interaction with a user. The processor 1001 and memory 1005 in the electronic device of the present invention may be provided in a decoding device or an encoding device, and the electronic device calls the decoding program and / or encoding program stored in the memory 1005 via the processor 1001 and executes the image decoding method or image encoding method according to an embodiment of the present invention.

[0022] An embodiment of the present invention provides an image decoding method, and please refer to FIG. 2, which is a schematic flowchart of a first embodiment of the image decoding method of the present invention.

[0023] In this embodiment, the image decoding method includes the following steps:

[0024] Step S10: Decode the image bitstream and determine the feature reconstruction values ​​corresponding to the current image features obtained by decoding.

[0025] In addition, the entity that executes this embodiment may be a decoding device when performing encoding processing on image data, and the decoding device may be an electronic device such as a personal computer or a server, or of course, it may be another device that can realize the same or similar functions.This embodiment is not limited to this, and in this embodiment and each of the following embodiments, the image decoding method of the present invention will be explained using a decoding device as an example.

[0026] In an image encoding process, the encoding device typically decodes the encoded image bitstream after encoding is completed, and determines whether the parameters used during encoding need to be adjusted based on the image quality of the image obtained by decoding, so the entity that performs this embodiment may be the encoding device.

[0027] The image bitstream may be a bitstream generated after the encoding device performs an encoding process on image data that needs to be compression-encoded. When the decoding device decodes the image bitstream, image features of the encoded image data are extracted from the image bitstream, and the image features currently obtained in the decoding process are current image features. The feature reconstruction values ​​may be image features obtained after feature reconstruction is performed on the current image features in the decoding process.

[0028] In a specific processing process, when encoding image data, the encoding device may divide the image data into one image block for processing, or of course, may divide the image data into multiple image blocks for processing, and this embodiment is not limited thereto.

[0029] Technical terms related to encoding or decoding an image include JPEG (Joint Photographic Experts Group), JPEG-AI (Joint Photographic Experts Group Artificial Intelligence), entropy encoding, neural network (NN), convolutional neural network (CNN), feature, rate-distortion optimized, etc., which will be explained here.

[0030] JPEG (Joint Photographic Experts Group) is a standard for continuous-tone still image compression. Its file extensions are .jpg and .jpeg, and it is the most common image file format. It primarily employs a joint coding method consisting of predictive coding (Differential Pulse Code Modulation, DPCM), discrete cosine transform (DCT), and entropy coding to remove redundant image and color data. It is a lossy compression format, compressing images into small storage spaces without damaging the image data to a certain extent. In particular, excessive compression rates can degrade the quality of the decompressed image, making excessive compression rates undesirable when high-quality images are desired.

[0031] The scope of JPEG AI is to create a machine-learning-based image coding standard that provides a single-stream, compact, compressed-domain representation, i.e., human-visible, significantly improving compression efficiency over commonly used image coding standards at the same subjective quality and effectively improving performance in image processing and computer vision tasks. JPEG AI supports a wide range of applications, including cloud storage, vision surveillance, autonomous driving, image collection, storage and management, real-time monitoring of vision data, and media distribution. The goal is to design a coding solution that significantly improves compression efficiency over commonly used coding standards at the same subjective quality and provides effective compressed-domain processing for machine learning-based image processing and computer vision tasks. Other key requirements include hardware / software-friendly encoding and decoding, support for 8-bit and 10-bit depth, and efficient encoding and progressive decoding for images with text and graphics.

[0032] Entropy coding is a coding method that follows the entropy principle to ensure that no information is lost during the coding process. Information entropy is the average amount of information (degree of uncertainty) in the information source. Common entropy coding methods include Shannon coding, Huffman coding, and arithmetic coding.

[0033] The neural network of the present invention refers to an artificial neural network, not a biological neural network. A neural network is a computational model composed of a large number of interconnected nodes (also called neurons). In an artificial neural network, neuron processing units can represent different objects, such as features, alphabets, concepts, or several meaningful abstract modes. The processing units in the network are divided into three types: input units, output units, and hidden units. Input units receive external signals and data, output units realize the output of system processing results, and hidden units are units located between the input units and output units that cannot be observed from outside the system. The connection weights between neurons reflect the connection strength between units, and information representation and processing are reflected in the connection relationships of the network processing units. An artificial neural network is a non-programmed, brain-like information processing method. Its essence is to obtain parallel, distributed information processing functions through network transformations and dynamic behavior, and to imitate the information processing functions of the human brain and nervous system to different degrees and levels. Currently, commonly used neural networks in the video processing field include CNN, RNN, and fully connected network (FCN).

[0034] A convolutional neural network (CNN) is a feedforward neural network and one of the most representative network structures in deep learning technology. Its artificial neurons can respond to peripheral units within a limited coverage area, demonstrating excellent performance in large-scale image processing. Generally, the basic structure of a CNN includes two layers: a feature extraction layer (also called a convolutional layer), in which each neuron's input is connected to the local receptive field of the previous layer to extract local features. Once the local feature is extracted, its positional relationship with other features is also determined. The other layer is a feature mapping layer (also called an activation layer). Each computational layer of the network consists of multiple feature mappings, each of which is a plane, with all neurons in the plane having equal weights. The feature mapping structure may use a sigmoid function, ReLU function, Leaky-ReLU function, PReLU function, or GDN (Generalized Difference Network) function as the activation function for the convolutional network. Furthermore, because neurons in a single mapping plane share weights, the number of free parameters in the network is reduced. One of the advantages of CNN compared to traditional image processing algorithms is that it avoids complex image preprocessing processes (such as extracting artificial features) and can directly input the original image for end-to-end learning.One of the advantages of CNN compared to traditional neural networks is that traditional neural networks all use a fully connected method, meaning that all neurons from the input layer to the hidden layer are fully connected, resulting in a huge number of parameters and making network training time-consuming and difficult, whereas CNN avoids this difficulty by using methods such as local connections and weight value sharing.

[0035] The feature according to the present invention is a three-dimensional feature matrix of C×W×H (as shown in FIG. 3, which is a schematic diagram of the matrix structure of this embodiment). C represents the number of channels, H represents the feature height, and W represents the feature width. The feature matrix can be the input or output of a neural network.

[0036] Coding efficiency is evaluated using two metrics: bit rate and peak signal-to-noise ratio (PSNR). The smaller the bitstream, the greater the compression rate; the higher the PSNR, the better the image coding efficiency. When selecting a mode, the discriminant essentially evaluates both metrics. The cost corresponding to the mode is: J(mode) = D + λ * R. Here, D represents distortion, typically evaluated using the sum of squared errors (SSE), where SSE refers to the root mean square of the difference between the reconstructed block and the source image; λ is the Lagrange multiplier; and R is the actual number of bits required to encode the image block in that mode, including the sum of bits required to encode mode information, motion information, residuals, etc. When selecting a mode, using the rate-distortion optimization (RDO) principle to compare and determine the coding mode usually ensures optimal coding performance.

[0037] In one possible embodiment of the present invention, the value of the image feature may be large and possibly complex. In order to improve the coding efficiency, the coding device can perform residual calculation on the image feature when coding the image feature, and then code the obtained residual data. When decoding the image data, a corresponding restoration process needs to be performed to obtain a feature reconstruction value close to the image feature before coding. In this case, the step S10 of this embodiment is as follows: decoding the image bitstream and determining residual reconstruction values ​​corresponding to the current image features obtained by the decoding; performing prediction based on the feature reconstruction values ​​of the reconstructed features to obtain predicted feature values; and determining a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value.

[0038] In addition, when the encoding device calculates and then encodes the residual data, when the encoding device decodes the image bitstream, it can obtain only the residual reconstruction value corresponding to the current image feature. The reconstructed feature may be a part of the image feature for which feature restoration has been completed, and performing prediction based on the feature reconstruction value of the reconstructed feature to obtain the predicted feature value may be performing prediction based on the feature reconstruction value of the reconstructed feature using an average prediction network to obtain the predicted feature value.

[0039] For ease of understanding, the description will be made with reference to FIG. 4, which is a schematic flowchart of the image encoding and decoding of this embodiment. In the figure, Bitstream#1 is the auxiliary bitstream, and Bitstream#2 is the image bitstream. As shown in FIG. 4, the analysis and transformation network performs feature extraction on the original image block (i.e., the image data that the encoding device needs to perform encoding processing) to obtain the image feature y. At the same time, the hyperparameter encoding network calculates the auxiliary information z_hat. Then, the mean value prediction network calculates the predicted feature value mu. The encoding device performs residual processing to obtain the residual original value r of the current feature. Then, after performing residual processing and quantization (Q&AE), the obtained encoding residual coefficient r_coef is written into the image bitstream (Bitstream#2).

[0040] Then, when processing the image bitstream (Bitstream#2), the decoding device extracts the coding residual coefficients r_coef from the image bitstream, performs inverse quantization and residual reconstruction (AD&IQ) on the coding residual coefficients, obtains a residual reconstruction value r_hat corresponding to the current image feature, performs prediction based on the feature reconstruction value y_hat of the reconstructed feature using a mean prediction network to obtain a predicted feature value mu, determines the feature reconstruction value y_hat of the current image feature based on the predicted feature value and the residual reconstruction value, performs feature enhancement on the feature reconstruction value to obtain an enhanced feature value y_hat_en, and finally performs synthesis encoding on the enhanced feature value using a synthesis transformation network to obtain a reconstructed image block x_hat. Here, the parameters used in the Q&AE and AD&IQ processing processes are all obtained after the probabilistic hyperparameter decoding network processes the auxiliary information z_hat.

[0041] Here, the analysis transformation network, the hyperparameter encoding network, the probabilistic hyperparameter decoding network, and the synthesis transformation network may all be neural networks constructed based on deep learning.

[0042] In practical use, determining the feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value may be by adding the residual reconstruction value and the predicted feature value and determining the sum as the feature reconstruction value corresponding to the current image feature. When making prediction, the feature reconstruction value of a reconstructed feature close to the position of the current image feature may be used to make the prediction.

[0043] In one possible embodiment of the present invention, in order to further improve coding efficiency, in the process of encoding an image, the coding device may, after calculating residual data, perform operations such as residual processing or quantization processing on the residual data, and write the coded residual coefficients obtained by the processing into the image bitstream; in this case, determining the residual reconstruction value corresponding to the current image feature obtained by decoding may be: decoding the image bitstream, extracting the coded residual coefficients corresponding to the current image feature, and performing inverse quantization and residual reconstruction on the coded residual coefficients to obtain the residual reconstruction value corresponding to the current image feature.

[0044] In one possible embodiment of the present invention, the prediction may be performed using the enhanced feature values ​​of the reconstructed features. In this case, step S10 of this embodiment may be decoding the image bitstream and determining residual reconstruction values ​​corresponding to the current image features obtained by the decoding; performing prediction based on the enhanced feature values ​​of the reconstructed features to obtain predicted feature values; and determining a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value.

[0045] In some cases, when calculating residual values ​​during the process of encoding image data, the encoding side may perform prediction using the emphasis feature value of the reconstructed feature to determine a predicted feature value corresponding to the image feature currently being encoded, and then calculate the encoding residual coefficient of the image feature currently being encoded based on the predicted feature value.Therefore, in the decoding process, a similar method must be used to perform prediction based on the emphasis feature value of the reconstructed feature using an average prediction network to obtain a predicted feature value, and then add the residual reconstruction value and the predicted feature value to obtain a feature reconstruction value corresponding to the current image feature.

[0046] For ease of understanding, the description will be made with reference to FIG. 5, which is a schematic flowchart of the image encoding and decoding of this embodiment. As shown in FIG. 5, in the process of encoding and decoding image data, the processing process is basically similar to that of FIG. 4 above. However, the difference is that in the decoding process, after obtaining the feature reconstruction value of the current image feature, a first feature emphasis method is adopted to perform feature emphasis on the feature reconstruction value (i.e., feature emphasis 1 in the figure), and the obtained first emphasized feature value y_hat_en1 is input into the synthesis transformation network for synthesis transformation processing to obtain the reconstructed image block x_hat. At the same time, a second feature emphasis method is adopted to perform feature emphasis on the feature reconstruction value (i.e., feature emphasis 2 in the figure), and the obtained second emphasized feature value y_hat_en2 is input into the average prediction network. Thereafter, when processing other image features, the current image feature is taken as the reconstructed feature, and prediction is performed using the second emphasized feature value to calculate the predicted feature value.

[0047] Note that feature enhancement 2 is typically performed one row at a time (the rows may be diagonal, not necessarily vertical) to generate subsequent feature predictions as shown in Figure 5, and the features for the entire image (or the entire image block) cannot be completed simultaneously. When feature enhancement 1 is performed, the features for the entire image (or the entire image block) are usually already reconstructed, so the features for the entire image (or the entire image block) may be performed in parallel.

[0048] In actual use, the feature enhancement methods of the reconstructed feature enhancement value and the current image feature enhancement value may be the same or different. That is, the feature enhancement processes of the first feature enhancement method and the second feature enhancement method may be the same process, in which case y_hat_en1 is equal to y_hat_en2, and the decoding device only needs to decode one set of related syntax parameters from the image bitstream. Of course, according to actual needs, the first feature enhancement method and the second feature enhancement method may be set to different processes, in which case the decoding device only needs to decode two sets of syntax parameters from the image bitstream.

[0049] In a specific application, an administrator of the decoding device or encoding device may set a configuration parameter, which can skip the first feature enhancement mode and / or the second feature enhancement mode, i.e., directly set y_hat_en1 and / or y_hat_en2 to y_hat, and the configuration parameter may be set in the image bitstream. Of course, the configuration parameter can also be set in other ways, and this embodiment is not limited thereto.

[0050] In one possible embodiment of the present invention, the encoding device may pre-write a syntax flag bit into the image bitstream to indicate whether the first feature enhancement scheme and the second feature enhancement scheme use exactly the same parameters. In this case, the decoding device may determine whether the first feature enhancement scheme and the second feature enhancement scheme use exactly the same parameters based on the syntax flag bit read from the image bitstream. For example, the encoding device may use a 1-bit syntax flag bit, useSameParaFlag, to indicate whether the first feature enhancement scheme and the second feature enhancement scheme use exactly the same syntax parameters. When useSameParaFlag=1 is read by the decoding device, it indicates that the first feature enhancement scheme and the second feature enhancement scheme use exactly the same syntax parameters. When useSameParaFlag=0 is read by the decoding device, it indicates that the first feature enhancement scheme and the second feature enhancement scheme use syntax parameters that are not exactly the same.

[0051] In one possible embodiment of the present invention, the syntax parameters associated with the feature enhancement scheme may be as shown in the table below. [Table 1]

[0052] In one possible embodiment of the present invention, other parameters, such as blockSizeList[idx] and modeList[idx], may be added based on the syntax parameters in Table 1. The code length of blockSizeList[idx] may be 8 bits, and the semantic is the block size N×N. A value of 1 for N indicates that it is based on each pixel. Specifically, if N is greater than 1, a representative value of the N×N block is obtained using methods such as minimum, maximum, or average, and then a new representative value is obtained by enhancing the representative value using the feature enhancement method of the present invention. The enhancement value of the N×N block is then set to the new representative value. The code length of modeList[idx] may be 3 bits, and values ​​1 to 4 indicate that min, avg, max, and max pools are used for up- and down-sampling, respectively (the block size is not 1×1). A value of 5 indicates that one set of filters has two scales. In this case, the parameters are as shown in Table 2. [Table 2]

[0053] Step S20: Perform feature enhancement on the feature reconstruction value to obtain an enhanced feature value.

[0054] When encoding image data, image features undergo processing such as quantization, resulting in distortion of the features. Therefore, the image quality of the reconstructed image can be improved by performing feature enhancement on the feature reconstruction value and then reconstructing the image based on the obtained enhanced feature value.

[0055] Step S30: Perform synthesis transformation on the enhancement feature value to obtain a reconstructed image block.

[0056] Note that obtaining a reconstructed image block by performing synthesis encoding on the enhancement feature value may also be obtaining a reconstructed image block by performing synthesis transformation processing on the enhancement feature value using a pre-constructed synthesis transformation network, where the synthesis transformation network may be a network constructed based on deep learning or a neural network.

[0057] Note that when the encoding device divides the image data into only one image block during the encoding process and processes it, the obtained reconstructed image block is reconstructed image data corresponding to the image data. When the encoding device divides the image data into multiple image blocks during the encoding process and processes it, the obtained reconstructed image block is reconstructed image data corresponding to an image block in the image data.

[0058] In this embodiment, an image bitstream is decoded, a feature reconstruction value corresponding to the current image feature obtained by decoding is determined, feature enhancement is performed on the feature reconstruction value to obtain an enhanced feature value, and a synthesis transformation is performed on the enhanced feature value to obtain a reconstructed image block. Before image reconstruction, feature enhancement is performed on the feature reconstruction value, and then image reconstruction is performed based on the enhanced feature enhancement value, thereby reducing distortion of image features in processes such as quantization and improving the image quality of the reconstructed image.

[0059] Please refer to FIG. 6, which is a schematic flowchart of a second embodiment of the image decoding method of the present invention.

[0060] Based on the first embodiment, step S20 of the image decoding method of this embodiment includes the following steps.

[0061] Step S201: Obtain the feature standard deviation and standard deviation expression value corresponding to each matrix element in the current image feature.

[0062] The current image feature may be a three-dimensional feature matrix, and the three dimensions may be represented by c, i, j, where c is a channel identifier, i and j are the feature height and feature width, respectively. The feature standard deviation corresponding to a matrix element may be the standard deviation between the feature corresponding to the matrix element and the feature mean value, and the standard deviation expression value is an expression value for expressing whether the standard deviation corresponding to the matrix element is a large standard deviation.

[0063] The standard deviation expression value may be divided into a first type expression value and a second type expression value (for simplicity, they may be expressed as true and false, where true is the first type expression value and false is the second type expression value), and when the standard deviation expression value corresponding to a matrix element is the first type expression value, it indicates that the standard deviation corresponding to the matrix element is a large standard deviation, and when the standard deviation expression value corresponding to a matrix element is the second type expression value, it indicates that the standard deviation corresponding to the matrix element is not a large standard deviation.

[0064] Step S202: Determine a feature mask corresponding to each matrix element in the current image feature according to the feature standard deviation, the standard deviation expression value and a preset specified threshold.

[0065] In addition, the feature mask may be an expression value for identifying whether or not feature enhancement is performed on the feature reconstruction value corresponding to the matrix element, and the feature mask may be divided into a first type value and a second type value (for simplicity, it may be expressed as true and false, where true is the first type value and false is the second type value), and when the feature mask corresponding to the matrix element is the first type value, it indicates that feature enhancement needs to be performed on the feature reconstruction value corresponding to the matrix element, and when the feature mask corresponding to the matrix element is the second type value, it indicates that feature enhancement does not need to be performed on the feature reconstruction value corresponding to the matrix element.

[0066] In practice, a preset emphasis condition may be set in advance, and the feature mask of the matrix element may be set by determining whether the feature standard deviation and standard deviation expression value corresponding to the matrix element meet the preset emphasis condition. In this case, step S202 of this embodiment may be: If the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a preset emphasis condition, setting the feature mask corresponding to the matrix element to a first type value; If the feature standard deviation and the standard deviation expression value corresponding to the matrix element do not satisfy the preset emphasis condition, the feature mask corresponding to the matrix element may be set to a second type value.

[0067] In addition, if the feature standard deviation and standard deviation expression value corresponding to a matrix element satisfy the preset emphasis condition, the feature reconstruction value corresponding to the matrix element indicates that feature emphasis is required, and therefore the feature mask corresponding to the matrix element can be set to the first type value.

[0068] If the feature standard deviation and standard deviation expression value corresponding to a matrix element do not satisfy the preset enhancement condition, the feature reconstruction value corresponding to the matrix element indicates that feature enhancement is not required, and therefore the feature mask corresponding to the matrix element can be set to a second type value.

[0069] In a specific implementation, feature enhancement may be performed for a matrix element that has a large standard deviation and whose corresponding feature standard deviation is greater than a certain threshold, or for a matrix element that does not have a large standard deviation and whose corresponding feature standard deviation is less than a certain threshold, where the predetermined threshold may be preset by an administrator of the encoding device or the decoding device. In this case, if the feature standard deviation and standard deviation representation value corresponding to the matrix element in this embodiment meet the predetermined enhancement condition, before the step of setting the feature mask corresponding to the matrix element to a first type value, determining that the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a predetermined emphasis condition when the feature standard deviation corresponding to the matrix element is greater than a predetermined threshold and the standard deviation expression value is a first type expression value; Or, The method may further include determining that the feature standard deviation and the standard deviation representation value corresponding to the matrix element satisfy a predetermined emphasis condition when the feature standard deviation corresponding to the matrix element is smaller than a predetermined threshold and the standard deviation representation value is a second type representation value.

[0070] In actual use, at least one filter may be used to perform feature enhancement on the feature reconstruction value, and the enhancement of each filter is performed sequentially according to the order of the filters (the order of filter execution may be preset by an administrator of the encoding device or decoding device).

[0071] Here, different indexes (idx) may be set for different filters to distinguish them. When determining the feature mask, different filters can set different feature masks for the same matrix element. For example, for a filter whose index is idx, the corresponding feature mask is:

number

[0072] mask[idx,c,i,j] is a feature mask set by the filter with index idx for the matrix element with matrix coordinates (c,i,j), Threshold[idx] is a preset threshold corresponding to the filter with index idx, GreaterFlag[idx] is a standard deviation representation value set by the filter with index idx for the matrix element with matrix coordinates (c,i,j), and σ[c,i,j] is the feature standard deviation corresponding to the matrix element with matrix coordinates (c,i,j).

[0073] In a specific implementation, if mask[idx,c,i,j] is true (1 and 0 may be used instead of true and false, in which case mask[idx,c,i,j]=1), the filter with index idx represents feature enhancement for the matrix element with matrix coordinates (c,i,j). In this case, if there are multiple sets of filters, feature enhancement may be performed multiple times for matrix elements with the same matrix coordinates.

[0074] Step S203: Based on the feature mask, perform feature enhancement on the feature reconstruction value to obtain an enhanced feature value.

[0075] In addition, performing feature emphasis on the feature reconstruction value based on the feature mask to obtain an emphasized feature value may also mean performing feature emphasis on the feature reconstruction value corresponding to some matrix elements in the current image feature that require feature emphasis based on the feature mask to obtain an emphasized feature value.

[0076] In this embodiment, the feature standard deviation and standard deviation expression value corresponding to each matrix element in the current image feature are obtained, and a feature mask corresponding to each matrix element in the current image feature is determined based on the feature standard deviation, the standard deviation expression value, and a predetermined threshold value, and feature enhancement is performed on the feature reconstruction value based on the feature mask to obtain an enhanced feature value. The feature mask corresponding to each matrix element is preset based on the feature standard deviation, the standard deviation expression value, and a predetermined threshold value corresponding to each matrix element, and by setting the feature mask, some matrix elements in the feature matrix corresponding to the current image feature that need to be enhanced are marked, so that when feature enhancement is performed, some matrix elements that need to be enhanced can be quickly determined, improving processing efficiency.

[0077] Please refer to FIG. 7, which is a schematic flowchart of a third embodiment of the image decoding method of the present invention.

[0078] Based on the second embodiment, the step S203 of the image decoding method of this embodiment includes the following steps:

[0079] Step S2031: A matrix element whose corresponding feature mask has a first type value is set as a target matrix element.

[0080] In addition, if the feature mask corresponding to a matrix element is a first type value, it indicates that enhancement is required for the feature reconstruction value corresponding to the matrix element. Therefore, first, based on the feature mask, matrix elements in the three-dimensional matrix corresponding to the current image feature may be selected, and the matrix elements whose corresponding feature mask is a first type value may be set as target matrix elements.

[0081] Step S2032: Perform enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value.

[0082] When emphasizing the feature reconstruction values ​​corresponding to the target matrix elements, the emphasis may be performed using a different method to obtain the emphasized feature values.

[0083] In one possible embodiment of the present invention, when enhancing the feature reconstruction value corresponding to the target matrix element, the enhancement may be performed by combining a preset scaling factor with the feature reconstruction value corresponding to the target matrix element, the residual reconstruction value, and the predicted feature value. In this case, step S2032 of this embodiment can be: obtaining feature reconstruction values, residual reconstruction values, and predicted feature values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the residual reconstruction value, and determining a second emphasis value based on a second scaling factor and the predicted feature value; determining an emphasis feature value based on the feature reconstruction value, the first emphasis value, and the second emphasis value.

[0084] It should be noted that the first scaling factor and the second scaling factor may be preset scaling factors, and different filters may correspond to different first scaling factors and second scaling factors.

[0085] In practical use, the emphasized feature value may be determined based on the feature reconstruction value, the first emphasized value, and the second emphasized value according to the first feature emphasized formula.

[0086] The first feature emphasis formula is: y_hat_en[c,i,j]=y_hat[c,i,j]+mean_hat[c,i,j]*Scale2[idx]+residual_hat[c,i,j]*Scale1[idx].

[0087] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat_en[c,i,j] is the enhancement feature value corresponding to the target matrix element, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, residual_hat[c,i,j]*Scale1[idx] is the first enhancement value, mean_hat[c,i,j]*Scale2[idx] is the second enhancement value, mean_hat[c,i,j] is the predicted feature value corresponding to the target matrix element, residual_hat[c,i,j] is the residual reconstruction value corresponding to the target matrix element, Scale1[idx] is the first scaling factor, and Scale2[idx] is the second scaling factor.

[0088] In one possible embodiment of the present invention, for the same filter, different scaling factors may be set when channels corresponding to different target matrix elements are different. In this case, before the step of determining a first emphasis value based on the first scaling factor and the predicted feature value in this embodiment and determining a second emphasis value based on the second scaling factor and the residual reconstruction value, further: obtaining a feature channel corresponding to the target matrix element; and determining a first scaling factor and a second scaling factor based on the feature channel.

[0089] Note that obtaining the feature channel corresponding to the target matrix element may involve obtaining the matrix coordinates (c, i, j) of the target matrix element, extracting c therefrom, and determining the feature channel corresponding to the target matrix element based on the value of c.

[0090] In actual use, determining the first scaling factor and the second scaling factor based on the feature channel may involve obtaining a factor channel mapping table corresponding to the currently used filter and searching for corresponding first scaling factor and second scaling factor in the factor channel mapping table based on the feature channel, where the factor channel mapping table includes a mapping relationship between the feature channel and the scaling factor, and different feature channels may correspond to different first scaling factor and second scaling factor in the factor channel mapping table, and the factor channel mapping table may be preset by an administrator of the encoding device or the decoding device.

[0091] In this case, the emphasized feature value may be determined based on the feature reconstruction value, the first emphasized value, and the second emphasized value according to a second feature emphasized formula, wherein the second feature emphasized formula is y_hat_en[c,i,j]=y_hat[c,i,j]+mean_hat[c,i,j]*Scale2[c,idx]+residual_hat[c,i,j]*Scale1[c,idx].

[0092] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat_en[c,i,j] is the enhancement feature value corresponding to the target matrix element, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, residual_hat[c,i,j]*Scale1[c,idx] is the first enhancement value, mean_hat[c,i,j]*Scale2[c,idx] is the second enhancement value, mean_hat[c,i,j] is the predicted feature value corresponding to the target matrix element, residual_hat[c,i,j] is the residual reconstruction value corresponding to the target matrix element, Scale1[c,idx] is the first scaling factor corresponding to the feature channel of the target matrix element, and Scale2[c,idx] is the second scaling factor corresponding to the feature channel of the target matrix element.

[0093] The administrator of the encoding device or decoding device may preset a control switch corresponding to each channel, and can turn off the emphasis on the feature of a certain channel by turning off the corresponding control switch, and of course, can turn off the emphasis on the feature of a certain channel by modifying the scaling factor corresponding to that channel in the factor-channel mapping table to 0.

[0094] In one possible embodiment of the present invention, when enhancing the feature reconstruction value corresponding to the target matrix element, the enhancement may be performed by combining a preset scaling factor with the feature reconstruction value corresponding to the target matrix element and the predicted feature value. In this case, step S2032 of this embodiment may be: obtaining feature reconstruction values ​​and predicted feature values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the predicted feature value; determining an emphasis feature value based on the first emphasis value and the second emphasis value.

[0095] The first and second scaling factors may be preset scaling factors, and different filters may correspond to different first and second scaling factors. Also, different first and second scaling factors may be set for the same filter according to different feature channels of the target matrix elements.

[0096] In actual use, the emphasized feature value may be determined based on the first emphasized value and the second emphasized value according to a third feature emphasized formula, which may be: y_hat_en[c,i,j]=y_hat[c,i,j]*Scale1[idx]+mean_hat[c,i,j]*Scale2[idx].

[0097] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat_en[c,i,j] is the enhancement feature value, y_hat[c,i,j]*Scale1[idx] is the first enhancement value, mean_hat[c,i,j]*Scale2[idx] is the second enhancement value, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, mean_hat[c,i,j] is the predicted feature value corresponding to the target matrix element, Scale1[idx] is the first scaling factor, and Scale2[idx] is the second scaling factor.

[0098] In one possible embodiment of the present invention, when enhancing the feature reconstruction value corresponding to the target matrix element, the enhancement may be performed by combining a preset scaling factor with the feature reconstruction value corresponding to the target matrix element and the residual reconstruction value. In this case, step S2032 of this embodiment can be: obtaining feature reconstruction values ​​and residual reconstruction values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the residual reconstruction value; determining an emphasis feature value based on the first emphasis value and the second emphasis value.

[0099] The first and second scaling factors may be preset scaling factors, and different filters may correspond to different first and second scaling factors. Also, different first and second scaling factors may be set for the same filter according to different feature channels of the target matrix elements.

[0100] In practical use, the emphasized feature value may be determined based on the first emphasized value and the second emphasized value according to a fourth feature emphasized formula, which may be: y_hat_en[c,i,j]=y_hat[c,i,j]*Scale1[idx]+residual_hat[c,i,j]*Scale2[idx].

[0101] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat_en[c,i,j] is the enhancement feature value, y_hat[c,i,j]*Scale1[idx] is the first enhancement value, residual_hat[c,i,j]*Scale2[idx] is the second enhancement value, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, residual_hat[c,i,j] is the residual reconstruction value corresponding to the target matrix element, Scale1[idx] is the first scaling factor, and Scale2[idx] is the second scaling factor.

[0102] In one possible embodiment of the present invention, when enhancing the feature reconstruction value corresponding to the target matrix element, the enhancement may be performed by combining a preset scaling factor with the feature reconstruction value and feature standard deviation corresponding to the target matrix element. In this case, step S2032 of this embodiment may be: obtaining feature reconstruction values ​​and feature standard deviations corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the feature standard deviation; determining an emphasis feature value based on the first emphasis value and the second emphasis value.

[0103] The first and second scaling factors may be preset scaling factors, and different filters may correspond to different first and second scaling factors. Also, different first and second scaling factors may be set for the same filter according to different feature channels of the target matrix elements.

[0104] In actual use, the emphasized feature value may be determined based on the first emphasized value and the second emphasized value according to a fifth feature emphasized formula, which may be: y_hat_en[c,i,j]=y_hat[c,i,j]*Scale1[idx]+σ[c,i,j]*Scale2[idx].

[0105] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat_en[c,i,j] is the enhancement feature value, y_hat[c,i,j]*Scale1[idx] is the first enhancement value, σ[c,i,j]*Scale2[idx] is the second enhancement value, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, σ[c,i,j] is the feature standard deviation corresponding to the target matrix element, Scale1[idx] is the first scaling factor, and Scale2[idx] is the second scaling factor.

[0106] In one possible embodiment of the present invention, the feature reconstruction values ​​include reconstruction values ​​corresponding to different components, such as luminance reconstruction values ​​and chrominance reconstruction values, and correspondingly, the enhancement feature values ​​may include enhancement values ​​of different components, such as luminance enhancement feature values ​​and chrominance enhancement feature values. When enhancing the feature reconstruction values ​​corresponding to the target matrix elements, the feature enhancement processes of different components may be independent of each other and do not affect each other. In this case, step S2032 in this embodiment can be: performing enhancement on the chromaticity reconstruction values ​​corresponding to the target matrix elements to obtain chromaticity enhancement feature values; performing enhancement on the luminance reconstruction values ​​corresponding to the target matrix elements to obtain luminance enhancement feature values.

[0107] When performing feature enhancement, different parameters may be set for enhancing different components in the same filter, for example, different scaling factors may be set.

[0108] In one possible embodiment of the present invention, secondary enhancement may be performed on the enhancement feature value of a certain component using the feature reconstruction value of another component. For example, when secondary enhancement is performed on the chrominance enhancement feature value using the luminance reconstruction value, after the step of obtaining the chrominance enhancement feature value by performing enhancement on the chrominance reconstruction value corresponding to the target matrix element in this embodiment, the following step may be performed: determining a first enhancement value based on the luminance reconstruction value and a first scaling factor; The method may further include a step of performing a secondary emphasis on the chromaticity emphasis feature value based on the first emphasis value.

[0109] It should be noted that the first scaling factor may be a preset scaling factor, and different filters may correspond to different first scaling factors.

[0110] In practice, a sixth feature emphasis formula may be used to perform secondary emphasis on the chromaticity emphasis feature value based on the first emphasis value. It may be y_hat_chroma_en2[c,i,j]=y_hat_chroma_en[c,i,j]+Scale1[idx]*y_hat_luma[c,i,j].

[0111] In the formula, y_hat_chroma_en2[c,i,j] is the chromaticity emphasis feature value after secondary emphasis, y_hat_chroma_en[c,i,j] is the chromaticity emphasis feature value, Scale1[idx]*y_hat_luma[c,i,j] is the first emphasis value, Scale1[idx] is the first scaling factor, and y_hat_luma[c,i,j] is the luminance reconstruction value corresponding to the target matrix element.

[0112] Of course, in a specific implementation, the luminance enhancement feature value may be subjected to secondary enhancement according to the chromaticity reconstruction value corresponding to the target matrix element.

[0113] In one possible embodiment of the present invention, the step of determining a first enhancement value based on the luminance reconstruction value and a first scaling factor in this embodiment, in order to clarify whether the decoding device needs to perform secondary enhancement, comprises: extracting component indication parameters from the image bitstream; If the component indication parameter is a chromaticity emphasis type parameter, determining a first emphasis value based on the luminance reconstruction value and a first scaling factor.

[0114] The component instruction parameter may be an instruction parameter for instructing whether or not to perform secondary emphasis and for which component emphasis feature values ​​secondary emphasis is required. For example, the value of the component instruction parameter may be 0 to 3, where 0 indicates that secondary emphasis is not required, 1 indicates that secondary emphasis is required for chromaticity emphasis feature values, 2 indicates that secondary emphasis is required for luminance emphasis feature values, and 3 indicates that secondary emphasis is required for both luminance emphasis feature values ​​and chromaticity emphasis feature values.

[0115] In addition, if the component instruction parameter is a chromaticity emphasis type parameter, it indicates that secondary emphasis needs to be performed and the component that needs secondary emphasis is a chromaticity emphasis feature value. In this case, a first emphasis value may be determined based on the luminance reconstruction value and the first scaling factor, and secondary emphasis may be performed on the chromaticity emphasis feature value based on the first emphasis value.

[0116] In one possible embodiment of the present invention, the filter may perform feature enhancement on all matrix elements in the current image feature using the same scaling factor, and in this case, may perform enhancement on the feature reconstruction values ​​corresponding to the target matrix elements based on a seventh feature enhancement formula, which may be: y_hat_en[c,i,j]=y_hat[c,i,j]*Scale1[idx].

[0117] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, y_hat_en[c,i,j] is the enhancement feature value, and Scale1[idx] is the scaling factor corresponding to the filter whose index is idx.

[0118] In one possible embodiment of the present invention, different scaling factors may be set to perform feature enhancement on matrix elements of different channels in the current image feature (of course, it is also allowed that the scaling factors of some of the same channels are the same), and in this case, enhancement may be performed on the feature reconstruction value corresponding to the target matrix element according to an eighth feature enhancement formula, which is: y_hat_en[c,i,j]=y_hat[c,i,j]*Scale1[c].

[0119] In the formula, (c,i,j) is the matrix coordinate of the target matrix element, idx is the filter index, y_hat[c,i,j] is the feature reconstruction value corresponding to the target matrix element, y_hat_en[c,i,j] is the enhancement feature value, and Scale1[c] is the scaling factor corresponding to the c channel.

[0120] In this embodiment, a matrix element whose corresponding feature mask is a first type value is set as a target matrix element, and the feature reconstruction value corresponding to the target matrix element is enhanced to obtain an enhanced feature value. First, a matrix element that needs to be subjected to some feature enhancement based on the feature mask is marked as a target matrix element, and then the feature reconstruction value of the target matrix element is enhanced, thereby reducing the number of matrix elements that need to be processed in the processing process and improving the execution efficiency of the image decoding method.

[0121] Please refer to FIG. 8, which is a schematic flowchart of a fourth embodiment of the image decoding method of the present invention.

[0122] Based on the first embodiment, step S20 of the image decoding method of this embodiment includes the following steps.

[0123] Step S201': Extract syntax application section parameters from the image bitstream, and obtain feature location information corresponding to the current image feature.

[0124] Note that, when performing feature enhancement on image features of an entire image block, if the same enhancement method is used for all feature enhancements (i.e., the same syntax parameters are used in the enhancement process), the parameters used in the enhancement process (e.g., the first scaling factor and the second scaling factor) may be optimal for only some image features in the image block and not optimal for other subsequent image features, which may even have a negative effect. To avoid this phenomenon, the encoding device may set multiple sets of syntax parameters when performing encoding, and use different syntax parameters to enhance image features at different positions in the image block.

[0125] In actual use, the syntax application interval parameter may be a parameter for indicating the application range corresponding to each set of syntax parameters, and the feature position information corresponding to the current image feature may include information such as the position of the current image feature in the image block, for example, the row number, the column number, etc.

[0126] Step S202': Determine emphasis syntax parameters based on the feature location information and the syntax application interval parameters.

[0127] In addition, determining the emphasis syntax parameter based on the feature location information and the syntax application section parameter may involve determining the application range of the syntax corresponding to each syntax parameter based on the syntax application section parameter, comparing the feature location information with the application range corresponding to each syntax parameter, determining the application range in which the feature location information is located, and setting the syntax parameter corresponding to the application range in which the feature location information is located as the emphasis syntax parameter.

[0128] In practice, two sets of syntax parameters, namely, first and second syntax parameters, may be set, and one syntax application interval parameter may be set to distinguish which parts of image features the first and second syntax parameters are used to enhance, respectively. Here, the syntax application interval parameter may limit the application range of the syntax parameters in dimensions such as channel range, row, column, and diagonal row.

[0129] For example, if there are two sets of syntax parameters set, namely the first syntax parameter and the second syntax parameter, and the syntax application section parameter appliedLineNum=K extracted from the bitstream at this time, the image features whose feature position information is in columns 1 to K (which may be rows or diagonal columns) are enhanced using the first syntax parameter, and the feature position information is enhanced from the Kth to the last columns using the second syntax parameter.

[0130] Of course, in actual use, two or more sets of syntax parameters may be set. In this case, one syntax application interval parameter may be set for each set of syntax parameters, and the application range of each set of syntax parameters may be determined based on the syntax application interval parameter.

[0131] For example, if a total of N sets of syntax parameters are set, there will also be N syntax application section parameters, which may be expressed as applyLineNum(i) (i = 1 to N), where applyLineNum(i) is the syntax application section parameter corresponding to the i-th set of syntax parameters. In this case, the application range of the first set of syntax parameters is the image features of columns 1 to applyLineNum(1) (which may be rows or diagonal columns), the application range of the second set of syntax parameters is the image features of columns applyLineNum(1) + 1 to applyLineNum(1) + applyLineNum(2) (which may be rows or diagonal columns), and the application range of the third set of syntax parameters is the image features of columns applyLineNum(2) + 1 to applyLineNum(2) + applyLineNum(3) (which may be rows or diagonal columns), and so on.

[0132] While it is relatively easy to understand which row or column an ​​image feature is located in, diagonal columns are relatively complicated. For ease of understanding, the description will be made with reference to FIG. 9 , but the present solution is not limited thereto. FIG. 9 is a schematic diagram of the feature reconstruction order of the present invention. As shown in FIG. 9 , when performing feature reconstruction, a reconstruction order is adopted in which reconstruction is performed in diagonal columns from the top left to the bottom right. In FIG. 9 , Row represents the row in which the image feature is located, Column represents the column in which the image feature is located, the dotted circle represents the reconstructed feature (samples that have already been processed), the black circle represents the image feature currently being reconstructed (current sample), and T represents the feature offset (wave) for each reconstruction. FIG. 9 shows that reconstruction is currently being performed on the image feature in the eighth diagonal column. At this time, the diagonal column in which the current image feature is located can be determined based on the feature position information (i.e., the row and column) of the current image feature. Then, based on the diagonal column in which the current image feature is located, a set of syntax parameters to be used for feature enhancement can be determined.

[0133] Step S203': Perform feature enhancement on the feature reconstruction value based on the enhancement syntax parameters to obtain an enhanced feature value.

[0134] Note that performing feature enhancement on the feature reconstruction value based on the enhancement syntax parameters to obtain the enhancement feature value may also be equivalent to performing feature enhancement using the enhancement syntax parameters and adopting a feature enhancement method provided in any of the embodiments of the image decoding method described above, and the explanation thereof will be omitted here.

[0135] In this embodiment, syntax application section parameters are extracted from the image bitstream, feature location information corresponding to the current image feature is obtained, enhancement syntax parameters are determined based on the feature location information and the syntax application section parameters, and feature enhancement is performed on the feature reconstruction value based on the enhancement syntax parameters to obtain an enhancement feature value. When performing feature enhancement, the enhancement syntax parameters specifically used for enhancement are determined based on the feature location information of the current image feature and the syntax application section parameters extracted from the image bitstream, so that different syntax parameters can be applied when performing feature enhancement on image features at different positions, thereby ensuring the best possible feature enhancement effect.

[0136] Please refer to FIG. 10, which is a schematic flowchart of a first embodiment of the image encoding method of the present invention.

[0137] In this embodiment, the image encoding method includes the following steps:

[0138] Step S910: Feature extraction is performed on the image block to be coded, and the extracted features are set as the current image features.

[0139] The entity that executes this embodiment may be the encoding device, which may be an electronic device such as a personal computer or a server, or may be another device that can realize the same or similar functions; this embodiment is not limited to this, and in this embodiment and each of the following embodiments, the image decoding method of the present invention will be described using an encoding device as an example.

[0140] The image block to be coded may be an image block obtained by dividing the image data that needs to be coded, or the image data may be divided into only one image block, or the image data may be divided into multiple image blocks.

[0141] In practical use, feature extraction may be performed on an image block to be coded and the extracted features may be used as the current image features, or feature extraction may be performed on an image block to be coded using an analysis and transformation network and the currently extracted image features may be used as the current image features.

[0142] Step S920: Perform prediction based on the feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values.

[0143] In practical use, performing prediction based on the feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values ​​may also be performing feature prediction based on the feature reconstruction values ​​corresponding to the reconstructed features using an average prediction network to obtain predicted feature values. Here, when performing prediction using an average prediction network, auxiliary information may be calculated using a hyperparameter coding network and input into the average prediction network, and prediction may be performed by combining the feature reconstruction values ​​corresponding to the reconstructed features and the auxiliary information in the average prediction network.

[0144] Step S930: Determine the coding residual coefficients corresponding to the current image features based on the predicted feature values.

[0145] In practical use, the predicted feature value may be subtracted from the feature value of the current image feature to obtain the residual original value, and the residual original value may be subjected to residual processing and quantization processing to obtain the coded residual coefficient.

[0146] Step S940: Write the coded residual coefficients into an image bitstream corresponding to the image block to be coded.

[0147] In addition, after the coding residual coefficients are written into the image bitstream corresponding to the image block to be coded, when image decoding is required, the coding residual coefficients can be directly read from the image bitstream, and the coding residual coefficients can be subjected to inverse quantization processing and residual restoration processing to obtain the residual reconstruction value. Finally, the predicted feature value of the mean value prediction network and the residual reconstruction value can be combined to determine the feature reconstruction value corresponding to the image block to be coded, and image reconstruction can be performed.

[0148] In actual use, the encoding device may perform operations such as parameter calculation, parameter setting, syntax setting, flag bit setting, etc., and specific embodiments can be obtained by referring to the content of any of the examples of the above image decoding method.

[0149] In this embodiment, feature extraction is performed on an image block to be coded, the extracted feature is taken as a current image feature, prediction is performed based on feature reconstruction values ​​corresponding to the reconstructed feature to obtain predicted feature values, coding residual coefficients corresponding to the current image feature are determined based on the predicted feature values, and the coding residual coefficients are written into an image bitstream corresponding to the image block to be coded. During coding, the coding residual coefficients of the current image feature are calculated and written into the image bitstream, thereby reducing the coding efficiency of coding image data.

[0150] Please refer to FIG. 11, which is a schematic flowchart of a second embodiment of the image coding method of the present invention.

[0151] Based on the first embodiment of the image coding method, this embodiment may further include the following steps after step S940.

[0152] Step S950: Image decoding is performed on the image bitstream corresponding to the image block to be coded to obtain a reconstructed image block.

[0153] In addition, after completing the encoding of image data, the encoding device needs to verify the encoding efficiency to ensure that the encoding efficiency is high. In this case, the encoding device may perform image decoding on the image bitstream corresponding to the image block to be encoded to obtain a reconstructed image block.

[0154] Step S960: Determine an image coding efficiency based on the reconstructed image block and the image block to be coded.

[0155] In practical use, determining the image coding efficiency based on the reconstructed image block and the image block to be coded may be determining the image coding efficiency by comparing the reconstructed image block with the image block to be coded and calculating the corresponding bit rate and PSNR based on the comparison result.

[0156] In one possible embodiment of the present invention, a preset efficiency threshold is set in advance, and after obtaining the image coding efficiency, the image coding efficiency is compared with the preset efficiency threshold. If the image coding efficiency is smaller than the preset efficiency threshold, it indicates that the image coding efficiency at this time is low. In this case, the parameters in the network or model used in the coding process can be adjusted to improve the image coding efficiency.

[0157] In practical use, when performing image decoding on an image bitstream corresponding to an image block to be coded, the image decoding method provided in any of the above image decoding method embodiments of the present invention may be used, and this embodiment is not limited thereto. The coding device may perform operations such as parameter calculation, parameter setting, syntax setting, and flag bit setting, and specific embodiments can be obtained by referring to the content of any of the above image coding method embodiments.

[0158] In this embodiment, image decoding is performed on an image bit stream corresponding to the image block to be coded to obtain a reconstructed image block, and the image coding efficiency is determined based on the reconstructed image block and the image block to be coded. After coding is completed and the image bit stream is obtained, image decoding is performed on the image bit stream, and the reconstructed image block obtained by image decoding is compared with the image block to be coded to determine the image coding efficiency. Therefore, parameters in each network used in the coding process can be adjusted based on the image coding efficiency, thereby improving the image coding efficiency.

[0159] Furthermore, an embodiment of the present invention further provides a storage medium on which an image decoding program and / or an image encoding program is stored, and when the image decoding program is executed, the above-mentioned image decoding method is implemented, and when the image encoding program is executed, the above-mentioned image encoding method is implemented.

[0160] Please refer to FIG. 12, which is a structural block diagram of a first embodiment of an image decoding device according to the present invention.

[0161] As shown in FIG. 12, an image decoding device provided by an embodiment of the present invention includes: a bitstream decoding module 10 for decoding the image bitstream and determining feature reconstruction values ​​corresponding to the current image features obtained by the decoding; a feature enhancement module 20 for performing feature enhancement on the feature reconstruction values ​​to obtain enhanced feature values; and an image reconstruction module 30 for performing a synthesis transform on the enhancement feature values ​​to obtain a reconstructed image block.

[0162] In this embodiment, an image bitstream is decoded, a feature reconstruction value corresponding to the current image feature obtained by decoding is determined, feature enhancement is performed on the feature reconstruction value to obtain an enhanced feature value, and a synthesis transformation is performed on the enhanced feature value to obtain a reconstructed image block. Before image reconstruction, feature enhancement is performed on the feature reconstruction value, and then image reconstruction is performed based on the enhanced feature enhancement value, thereby reducing distortion of image features in processes such as quantization and improving the image quality of the reconstructed image.

[0163] In one possible embodiment of the present invention, the current image features are a three-dimensional feature matrix; The feature enhancement module 20 is further used to obtain a feature standard deviation and a standard deviation representation value corresponding to each matrix element in the current image feature, determine a feature mask corresponding to each matrix element in the current image feature based on the feature standard deviation, the standard deviation representation value and a predetermined specified threshold, and perform feature enhancement on the feature reconstruction value based on the feature mask to obtain an enhanced feature value.

[0164] In one possible embodiment of the present invention, the feature highlighting module 20 is further used to set the feature mask corresponding to the matrix element to a first type value if the feature standard deviation and standard deviation representation value corresponding to the matrix element meet the preset highlighting condition, and to set the feature mask corresponding to the matrix element to a second type value if the feature standard deviation and standard deviation representation value corresponding to the matrix element do not meet the preset highlighting condition.

[0165] In one possible embodiment of the present invention, the feature highlighting module 20 is further used to determine that the feature standard deviation and standard deviation representation value corresponding to the matrix element satisfy the predetermined highlighting condition if the feature standard deviation corresponding to the matrix element is greater than a predetermined threshold and the standard deviation representation value is a first type representation value, or to determine that the feature standard deviation and standard deviation representation value corresponding to the matrix element satisfy the predetermined highlighting condition if the feature standard deviation corresponding to the matrix element is smaller than a predetermined threshold and the standard deviation representation value is a second type representation value.

[0166] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to set a matrix element whose corresponding feature mask is a first type value as a target matrix element, and perform enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value.

[0167] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to obtain feature reconstruction values, residual reconstruction values, and predicted feature values ​​corresponding to the target matrix elements, determine a first enhancement value based on a first scaling factor and the residual reconstruction value, determine a second enhancement value based on a second scaling factor and the predicted feature value, and determine an enhancement feature value based on the feature reconstruction values, the first enhancement value, and the second enhancement value.

[0168] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to obtain feature channels corresponding to the target matrix elements, and determine first and second scaling factors based on the feature channels, where different feature channels correspond to different first and second scaling factors.

[0169] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to obtain feature reconstruction values ​​and predicted feature values ​​corresponding to the target matrix elements, determine a first enhancement value based on a first scaling factor and the feature reconstruction values, determine a second enhancement value based on a second scaling factor and the predicted feature values, and determine an enhancement feature value based on the first enhancement value and the second enhancement value.

[0170] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to obtain feature reconstruction values ​​and residual reconstruction values ​​corresponding to the target matrix elements, determine a first enhancement value based on a first scaling factor and the feature reconstruction values, determine a second enhancement value based on a second scaling factor and the residual reconstruction values, and determine an enhancement feature value based on the first enhancement value and the second enhancement value.

[0171] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to obtain a feature reconstruction value and a feature standard deviation corresponding to the target matrix element, determine a first enhancement value based on a first scaling factor and the feature reconstruction value, determine a second enhancement value based on a second scaling factor and the feature standard deviation, and determine an enhanced feature value based on the first enhancement value and the second enhancement value.

[0172] In one possible embodiment of the present invention, the feature reconstruction values ​​include luminance reconstruction values ​​and chrominance reconstruction values, and the enhancement feature values ​​include luminance enhancement feature values ​​and chrominance enhancement feature values; The feature enhancement module 20 is further used to perform enhancement on the chrominance reconstruction values ​​corresponding to the target matrix elements to obtain chrominance-enhanced feature values, and to perform enhancement on the luma reconstruction values ​​corresponding to the target matrix elements to obtain luma-enhanced feature values.

[0173] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to determine a first enhancement value based on the luminance reconstruction value and a first scaling factor, and to perform a secondary enhancement on the chrominance enhancement feature value based on the first enhancement value.

[0174] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to extract a component indication parameter from the image bitstream, and if the component indication parameter is a chrominance enhancement type parameter, to determine a first enhancement value based on the luma reconstruction value and a first scaling factor.

[0175] In one possible embodiment of the present invention, the bitstream decoding module 10 is further used to decode an image bitstream, determine a residual reconstruction value corresponding to a current image feature obtained by decoding, perform prediction based on the feature reconstruction values ​​of the reconstructed features to obtain a predicted feature value, and determine a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value.

[0176] In one possible embodiment of the present invention, the bitstream decoding module 10 is further used to perform prediction based on the enhancement feature value of the reconstructed feature to obtain a predicted feature value, and to determine a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value.

[0177] In one possible embodiment of the present invention, the feature enhancement schemes of the enhancement feature values ​​of the reconstructed features and the enhancement feature values ​​corresponding to the current image features are the same or different.

[0178] In one possible embodiment of the present invention, whether the feature enhancement schemes of the enhancement feature values ​​of the reconstructed features and the enhancement feature values ​​corresponding to the current image features are the same or different is determined by a syntax flag bit, which is read from the image bitstream.

[0179] In one possible embodiment of the present invention, the feature enhancement module 20 is further used to extract syntax application interval parameters from the image bitstream, obtain feature location information corresponding to the current image feature, determine enhancement syntax parameters based on the feature location information and the syntax application interval parameters, and perform feature enhancement on the feature reconstruction value based on the enhancement syntax parameters to obtain an enhancement feature value.

[0180] Please refer to FIG. 13, which is a structural block diagram of a first embodiment of an image encoding device according to the present invention.

[0181] As shown in FIG. 13, the image coding device provided by the embodiment of the present invention includes: a feature extraction module 110 for performing feature extraction on the image block to be coded and setting the extracted features as current image features; a feature prediction module 120 for performing prediction based on feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values; a residual calculation module 130 for determining coding residual coefficients corresponding to the current image features based on the predicted feature values; a parameter writing module 140 for writing the coded residual coefficients into an image bitstream corresponding to the image block to be coded.

[0182] In this embodiment, feature extraction is performed on an image block to be coded, the extracted feature is taken as a current image feature, prediction is performed based on feature reconstruction values ​​corresponding to the reconstructed feature to obtain predicted feature values, coding residual coefficients corresponding to the current image feature are determined based on the predicted feature values, and the coding residual coefficients are written into an image bitstream corresponding to the image block to be coded. During coding, the coding residual coefficients of the current image feature are calculated and written into the image bitstream, thereby reducing the coding efficiency of coding image data.

[0183] In one possible embodiment of the present invention, the parameter writing module 140 is further used to perform image decoding on an image bitstream corresponding to the image block to be encoded to obtain a reconstructed image block, and determine an image encoding efficiency based on the reconstructed image block and the image block to be encoded.

[0184] It should be noted that the above is merely an example and does not limit the technical solution of the present invention. In specific applications, those skilled in the art can set it as needed, and the present invention is not limited thereto.

[0185] It should be noted that the above-described flow is merely an example and does not limit the scope of protection of the present invention. In actual applications, those skilled in the art can select part or all of it according to actual needs to achieve the objective of the technical solution of this embodiment, and no limitation is made here.

[0186] Furthermore, for technical details not described in detail in this embodiment, reference may be made to the image decoding or image decoding method provided in any embodiment of the present invention, and the description will be omitted here.

[0187] It should be noted that, as used herein, the terms "comprise," "contain," or any other variation thereof, are intended to include a non-exclusive inclusion, whereby a process, method, article, or system that includes a set of elements not only includes those elements, but also includes other elements not expressly listed, or includes the inherent elements of such process, method, article, or system. Absent more limitations, an element qualified by "comprises a..." does not exclude the presence of additional identical elements in a process, method, article, or system that includes that element.

[0188] The numbers of the above embodiments of the present invention are for illustrative purposes only and do not represent the superiority of the embodiments.

[0189] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be realized by software and a necessary general-purpose hardware platform, and of course, can also be realized by hardware, and in many cases, the former is a more preferred embodiment. Based on this understanding, the technical solution of the present invention, or a part that contributes to the prior art, may be embodied in the form of a software product, and the computer software product is stored in a storage medium (e.g., a read-only memory (ROM) / RAM, a magnetic disk, an optical disk), and includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0190] The above are merely optional embodiments of the present invention and are not intended to limit the scope of the present invention. Any equivalent structure or equivalent process transformation made by utilizing the contents of the specification and drawings of the present invention, or any direct or indirect application to other related technical fields, are all equally included in the scope of protection of the present invention.

Claims

1. decoding the image bitstream and determining feature reconstruction values ​​corresponding to the current image features obtained by the decoding; performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value; and performing a synthesis transform on the enhancement feature values ​​to obtain a reconstructed image block. Image decoding method.

2. the current image features are a three-dimensional feature matrix; performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value, obtaining a feature standard deviation and a standard deviation representation value corresponding to each matrix element in the current image feature; determining a feature mask corresponding to each matrix element in the current image feature based on the feature standard deviation, the standard deviation representation value, and a predetermined threshold value; and performing feature enhancement on the feature reconstruction value based on the feature mask to obtain an enhanced feature value. The image decoding method according to claim 1 .

3. determining a feature mask corresponding to each matrix element in the current image feature based on the feature standard deviation, the standard deviation representation value, and a predetermined threshold value, If the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a preset emphasis condition, setting the feature mask corresponding to the matrix element to a first type value; If the feature standard deviation and the standard deviation expression value corresponding to the matrix element do not satisfy the preset emphasis condition, setting the feature mask corresponding to the matrix element to a second type value. The image decoding method according to claim 2 .

4. If the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a preset emphasis condition, before setting the feature mask corresponding to the matrix element to a first type value, further: determining that the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a predetermined emphasis condition when the feature standard deviation corresponding to the matrix element is greater than a predetermined threshold and the standard deviation expression value is a first type expression value; Or, determining that the feature standard deviation and the standard deviation expression value corresponding to the matrix element satisfy a predetermined emphasis condition when the feature standard deviation corresponding to the matrix element is smaller than a predetermined threshold value and the standard deviation expression value is a second type expression value; The image decoding method according to claim 3 .

5. performing feature enhancement on the feature reconstruction value based on the feature mask to obtain an enhanced feature value, determining a matrix element whose corresponding feature mask has a first type value as a target matrix element; performing enhancement on the feature reconstruction values ​​corresponding to the target matrix elements to obtain enhanced feature values; The image decoding method according to claim 2 .

6. The step of performing enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value includes: obtaining feature reconstruction values, residual reconstruction values, and predicted feature values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the residual reconstruction value, and determining a second emphasis value based on a second scaling factor and the predicted feature value; determining an emphasis feature value based on the feature reconstruction value, the first emphasis value, and the second emphasis value. The image decoding method according to claim 5 .

7. before determining a first emphasis value based on a first scaling factor and the predicted feature value and determining a second emphasis value based on a second scaling factor and the residual reconstruction value, further obtaining a feature channel corresponding to the target matrix element; determining a first scaling factor and a second scaling factor based on the feature channel; different feature channels correspond to different first and second scaling factors; The image decoding method according to claim 6 .

8. The step of performing enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value includes: obtaining feature reconstruction values ​​and predicted feature values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the predicted feature value; determining an emphasis feature value based on the first emphasis value and the second emphasis value; The image decoding method according to claim 5 .

9. The step of performing enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value includes: obtaining feature reconstruction values ​​and residual reconstruction values ​​corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the residual reconstruction value; determining an emphasis feature value based on the first emphasis value and the second emphasis value; The image decoding method according to claim 5 .

10. The step of performing enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value includes: obtaining feature reconstruction values ​​and feature standard deviations corresponding to the target matrix elements; determining a first emphasis value based on a first scaling factor and the feature reconstruction value, and determining a second emphasis value based on a second scaling factor and the feature standard deviation; determining an emphasis feature value based on the first emphasis value and the second emphasis value; The image decoding method according to claim 5 .

11. the feature reconstruction values ​​include luminance reconstruction values ​​and chrominance reconstruction values, and the enhancement feature values ​​include luminance enhancement feature values ​​and chrominance enhancement feature values; The step of performing enhancement on the feature reconstruction value corresponding to the target matrix element to obtain an enhanced feature value includes: performing enhancement on the chromaticity reconstruction values ​​corresponding to the target matrix elements to obtain chromaticity enhancement feature values; performing enhancement on the luminance reconstruction values ​​corresponding to the target matrix elements to obtain luminance-enhanced feature values; The image decoding method according to claim 5 .

12. After performing enhancement on the chromaticity reconstruction value corresponding to the target matrix element to obtain a chromaticity enhancement feature value, determining a first enhancement value based on the luminance reconstruction value and a first scaling factor; performing a secondary emphasis on the chromaticity emphasis feature value based on the first emphasis value; The image decoding method according to claim 11.

13. The step of determining a first enhancement value based on the luminance reconstruction value and a first scaling factor further comprises: extracting component indication parameters from the image bitstream; and if the component indication parameter is a chromaticity enhancement type parameter, determining a first enhancement value based on the luminance reconstruction value and a first scaling factor. The image decoding method according to claim 12.

14. The step of decoding the image bitstream and determining a feature reconstruction value corresponding to a current image feature obtained by decoding comprises: decoding the image bitstream and determining residual reconstruction values ​​corresponding to the current image features obtained by the decoding; performing prediction based on the feature reconstruction values ​​of the reconstructed features to obtain predicted feature values; determining a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value; The image decoding method according to any one of claims 1 to 13.

15. After decoding the image bitstream and determining the residual reconstruction value corresponding to the current image feature obtained by decoding, further: performing prediction based on the enhanced feature values ​​of the reconstructed features to obtain predicted feature values; determining a feature reconstruction value corresponding to the current image feature based on the residual reconstruction value and the predicted feature value; The image decoding method according to claim 14.

16. The feature enhancement methods of the enhancement feature values ​​of the reconstructed features and the enhancement feature values ​​corresponding to the current image features are the same or different. The image decoding method according to claim 14.

17. whether the feature enhancement schemes of the reconstructed feature and the enhancement feature value corresponding to the current image feature are the same or different are determined by a syntax flag bit, and the syntax flag bit is read from the image bitstream; The image decoding method according to claim 16.

18. performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value, extracting syntax application section parameters from the image bitstream to obtain feature location information corresponding to the current image feature; determining an emphasis syntax parameter based on the feature location information and the syntax application interval parameter; performing feature enhancement on the feature reconstruction value based on the enhancement syntax parameters to obtain an enhanced feature value; The image decoding method according to claim 14.

19. a bitstream decoding module for decoding the image bitstream and determining a feature reconstruction value corresponding to a current image feature obtained by the decoding; a feature enhancement module for performing feature enhancement on the feature reconstruction value to obtain an enhanced feature value; an image reconstruction module for performing a synthesis transform on the enhancement feature values ​​to obtain a reconstructed image block. Image decoding device.

20. performing feature extraction on the image block to be encoded and setting the extracted features as current image features; performing prediction based on feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values; determining coding residual coefficients corresponding to the current image features based on the predicted feature values; writing the coding residual coefficients into an image bitstream corresponding to the image block to be coded. Image encoding method.

21. After writing the coding residual coefficients into an image bitstream corresponding to the image block to be coded, further performing image decoding on an image bitstream corresponding to the image block to be coded to obtain a reconstructed image block; determining an image coding efficiency based on the reconstructed image block and the image block to be coded; 21. The image coding method according to claim 20.

22. a feature extraction module for performing feature extraction on the image block to be encoded and setting the extracted features as current image features; a feature prediction module for performing prediction based on feature reconstruction values ​​corresponding to the reconstructed features to obtain predicted feature values; a residual calculation module for determining coding residual coefficients corresponding to the current image feature based on the predicted feature values; a parameter writing module for writing the coding residual coefficients into an image bitstream corresponding to the image block to be coded. Image encoding device.

23. A decoding device including a processor, a memory, and a decoding program stored in the memory and executable by the processor, wherein when the decoding program is executed by the processor, the image decoding method according to any one of claims 1 to 18 is performed. Decryption device.

24. An encoding device including a processor, a memory, and a decoding program and / or an encoding program stored in the memory and executable by the processor, wherein when the decoding program is executed by the processor, the image decoding method according to any one of claims 1 to 18 is implemented, and when the encoding program is executed by the processor, the image encoding method according to any one of claims 19 to 20 is implemented. Encoding device.

25. A computer-readable storage medium storing an image decoding program and / or an image encoding program, wherein, when the image decoding program is executed, the image decoding method according to any one of claims 1 to 18 is implemented, and when the image encoding program is executed, the image encoding method according to any one of claims 19 to 20 is implemented. A computer-readable storage medium.

Images