Implicit multiple transform selection (MTS) using offline transform analysis

Abstract

In one implementation, we propose an implicit MTS method based on a mode-wise lookup table (LUT), where the LUT infers the best transform pair for a given block from its intra prediction mode and / or TU size. The LUT is trained offline using statistics collected from an exhaustive search on transforms performed in the encoder. This approach offers a flexible distribution of transform pairs between consecutive modes and allows combinations of other transforms with DCT2, which are not considered in the current ECM explicit MTS LUT. The method applies at both the encoder and decoder processes.

Description

IMPLICIT MULTIPLE TRANSFORM SELECTION (MTS) USING OFFLINE TRANSFORM ANALYSISTECHNICAL FIELD[1] The present embodiments generally relate to a method and an apparatus for transform coding in video encoding and decoding.BACKGROUND[2] To achieve high compression efficiency, image and video coding schemes usually employ prediction and transform to leverage spatial and temporal redundancy in the video content. Generally, intra or inter prediction is used to exploit the intra or inter picture correlation, then the differences between the original block and the predicted block, often denoted as prediction errors or prediction residuals, are transformed, quantized, and entropy coded. To reconstruct the video, the compressed data are decoded by inverse processes corresponding to the entropy coding, quantization, transform, and prediction.SUMMARY[3] According to one embodiment, a method of video decoding is presented, comprising: selecting a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; inverse transforming transform coefficients of said current block based on said single transform pair; obtaining residuals for said current block from said inverse transformed coefficients; and decoding said current block based on said residuals.[4] According to another embodiment, a method of video encoding is presented, comprising: selecting a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unitsizes; obtaining residuals for said current block; transforming said residuals for said current block to obtain transform coefficients of said current block based on said single transform pair; and encoding said current block based on said transform coefficients.[5] According to another embodiment, an apparatus for video decoding is presented, comprising at least a memory and one or more processors, wherein said one or more processors are configured to: select a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; inverse transform transform coefficients of said current block based on said single transform pair; obtain residuals for said current block from said inverse transformed coefficients; and decode said current block based on said residuals.[6] According to another embodiment, an apparatus for video encoding is presented, comprising at least a memory and one or more processors, wherein said one or more processors are configured to: select a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; obtain residuals for said current block; transform said residuals for said current block to obtain transform coefficients of said current block based on said single transform pair; and encode said current block based on said transform coefficients.[7] In one embodiment, said single transform pair is selected based on a look up table, wherein said look up table indicates a mapping between (1) said at least one of said intra prediction mode and said transform unit size and (2) said single transform pair.[8] In one embodiment, said plurality of transform pairs include a pair of transforms, which includes a DCT2 transform and another trigonometric transform.[9] In one embodiment, said plurality of intra prediction modes include all intra prediction modes.

[0010] In one embodiment, a subset of intra prediction modes is available at a parsing stage, andwherein said plurality of intra prediction modes only include intra prediction modes excluding said subset of intra prediction modes.

[0011] In one embodiment, said subset of intra prediction modes corresponds to at least one of PLANAR and MIP modes.

[0012] In one embodiment, said plurality of intra prediction modes include at least one intra prediction mode that is available at parsing.

[0013] In one embodiment, explicit signaling is used for a transform pair corresponding to a TU with an intra prediction mode of said subset of intra prediction modes.

[0014] In one embodiment, said plurality of transform unit sizes include all possible TU sizes.

[0015] One or more embodiments also provide a computer program comprising instructions which when executed by one or more processors cause the one or more processors to perform the encoding method or decoding method according to any of the embodiments described herein. One or more of the present embodiments also provide a computer readable storage medium having stored thereon instructions for encoding or decoding video data according to the methods described herein.

[0016] One or more embodiments also provide a computer readable storage medium having stored thereon video data generated according to the methods described above. One or more embodiments also provide a method and apparatus for transmitting or receiving the video data generated according to the methods described herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0017] FIG. 1 illustrates a block diagram of a system within which aspects of the present embodiments may be implemented.

[0018] FIG. 2 illustrates a block diagram of an embodiment of a video encoder.

[0019] FIG. 3 illustrates a block diagram of an embodiment of a video decoder.

[0020] FIG. 4 illustrates the intra prediction modes in VVC.

[0021] FIG. 5 illustrates that the ISP (Intra Sub -Partition) tool divides luma intra-predicted blocks vertically or horizontally into 2 sub-partitions depending on the block size dimensions.

[0022] FIG. 6 illustrates that the ISP tool divides luma intra-predicted blocks vertically or horizontally into 4 sub-partitions depending on the block size dimensions.

[0023] FIG. 7 illustrates an example of transform set selection in the Look Up Table (LUT) for a given TU size and intra mode.

[0024] FIG. 8 illustrates a process of implicit MTS using mode-wise LUT for the decoder, according to an embodiment.

[0025] FIG. 9 illustrates a hybrid explicit / implicit MTS method at the encoder, according to an embodiment.

[0026] FIG. 10 illustrates a hybrid explicit / implicit MTS method at the decoder, according to an embodiment.

[0027] FIG. 11 illustrates a training process to select transform pairs based on RD costs, according to an embodiment.

[0028] FIG. 12 illustrates an example of transform pair ranking based on the RD-cost.

[0029] FIG. 13 A and FIG. 13B illustrate the mode-wise LUT in the form of a heatmap, without and with noise applied on the statistics, respectively.

[0030] FIGs. 14A-14D illustrate the transform pair index for each intra mode and TU size.

[0031] FIG. 15 displays the BD-rate and complexity tradeoff of each approach in a 2D plane.DETAILED DESCRIPTION

[0032] FIG. 1 illustrates a block diagram of an example of a system in which various aspects and embodiments can be implemented. System 100 may be embodied as a device including the various components described below and is configured to perform one or more of the aspects described in this application. Examples of such devices, include, but are not limited to, various electronic devices such as personal computers, laptop computers, smartphones, tablet computers, digital multimedia settop boxes, digital television receivers, personal video recording systems, connected home appliances, and servers. Elements of system 100, singly or in combination, may be embodied in a single integrated circuit, multiple ICs, and / or discrete components. For example, in at least one embodiment, the processing and encoder / decoder elements of system 100 are distributed across multiple ICs and / or discrete components. In various embodiments, the system100 is communicatively coupled to other systems, or to other electronic devices, via, for example, a communications bus or through dedicated input and / or output ports. In various embodiments, the system 100 is configured to implement one or more of the aspects described in this application.

[0033] The system 100 includes at least one processor 110 configured to execute instructions loaded therein for implementing, for example, the various aspects described in this application. Processor 110 may include embedded memory, input output interface, and various other circuitries as known in the art. The system 100 includes at least one memory 120 (e.g., a volatile memory device, and / or a non-volatile memory device). System 100 includes a storage device 140, which may include non-volatile memory and / or volatile memory, including, but not limited to, EEPROM, ROM, PROM, RAM, DRAM, SRAM, flash, magnetic disk drive, and / or optical disk drive. The storage device 140 may include an internal storage device, an attached storage device, and / or a network accessible storage device, as non-limiting examples.

[0034] System 100 includes an encoder / decoder module 130 configured, for example, to process data to provide an encoded video or decoded video, and the encoder / decoder module 130 may include its own processor and memory. The encoder / decoder module 130 represents module(s) that may be included in a device to perform the encoding and / or decoding functions. As is known, a device may include one or both of the encoding and decoding modules. Additionally, encoder / decoder module 130 may be implemented as a separate element of system 100 or may be incorporated within processor 110 as a combination of hardware and software as known to those skilled in the art.

[0035] Program code to be loaded onto processor 110 or encoder / decoder 130 to perform the various aspects described in this application may be stored in storage device 140 and subsequently loaded onto memory 120 for execution by processor 110. In accordance with various embodiments, one or more of processor 110, memory 120, storage device 140, and encoder / decoder module 130 may store one or more of various items during the performance of the processes described in this application. Such stored items may include, but are not limited to, the input video, the decoded video or portions of the decoded video, the bitstream, matrices, variables, and intermediate or final results from the processing of equations, formulas, operations, and operational logic.

[0036] In several embodiments, memory inside of the processor 110 and / or the encoder / decoder module 130 is used to store instructions and to provide working memory for processing that isneeded during encoding or decoding. In other embodiments, however, a memory external to the processing device (for example, the processing device may be either the processor 110 or the encoder / decoder module 130) is used for one or more of these functions. The external memory may be the memory 120 and / or the storage device 140, for example, a dynamic volatile memory and / or a non-volatile flash memory. In several embodiments, an external non-volatile flash memory is used to store the operating system of a television. In at least one embodiment, a fast external dynamic volatile memory such as a RAM is used as working memory for video coding and decoding operations, such as for MPEG-2, HEVC, or VVC.

[0037] The input to the elements of system 100 may be provided through various input devices as indicated in block 105. Such input devices include, but are not limited to, (i) an RF portion that receives an RF signal transmitted, for example, over the air by a broadcaster, (ii) a Composite input terminal, (iii) a USB input terminal, and / or (iv) an HDMI input terminal.

[0038] In various embodiments, the input devices of block 105 have associated respective input processing elements as known in the art. For example, the RF portion may be associated with elements suitable for (i) selecting a desired frequency (also referred to as selecting a signal, or band-limiting a signal to a band of frequencies), (ii) down converting the selected signal, (iii) bandlimiting again to a narrower band of frequencies to select (for example) a signal frequency band which may be referred to as a channel in certain embodiments, (iv) demodulating the down converted and band-limited signal, (v) performing error correction, and (vi) demultiplexing to select the desired stream of data packets. The RF portion of various embodiments includes one or more elements to perform these functions, for example, frequency selectors, signal selectors, bandlimiters, channel selectors, filters, downconverters, demodulators, error correctors, and demultiplexers. The RF portion may include a tuner that performs various of these functions, including, for example, down converting the received signal to a lower frequency (for example, an intermediate frequency or a near-baseband frequency) or to baseband. In one set-top box embodiment, the RF portion and its associated input processing element receives an RF signal transmitted over a wired (for example, cable) medium, and performs frequency selection by filtering, down converting, and filtering again to a desired frequency band. Various embodiments rearrange the order of the above-described (and other) elements, remove some of these elements, and / or add other elements performing similar or different functions. Adding elements may includeinserting elements in between existing elements, for example, inserting amplifiers and an analog- to-digital converter. In various embodiments, the RF portion includes an antenna.

[0039] Additionally, the USB and / or HDMI terminals may include respective interface processors for connecting system 100 to other electronic devices across USB and / or HDMI connections. It is to be understood that various aspects of input processing, for example, Reed-Solomon error correction, may be implemented, for example, within a separate input processing IC or within processor 110 as necessary. Similarly, aspects of USB or HDMI interface processing may be implemented within separate interface ICs or within processor 110 as necessary. The demodulated, error corrected, and demultiplexed stream is provided to various processing elements, including, for example, processor 110, and encoder / decoder 130 operating in combination with the memory and storage elements to process the datastream as necessary for presentation on an output device.

[0040] Various elements of system 100 may be provided within an integrated housing, Within the integrated housing, the various elements may be interconnected and transmit data therebetween using suitable connection arrangement 115, for example, an internal bus as known in the art, including the I2C bus, wiring, and printed circuit boards.

[0041] The system 100 includes communication interface 150 that enables communication with other devices via communication channel 190. The communication interface 150 may include, but is not limited to, a transceiver configured to transmit and to receive data over communication channel 190. The communication interface 150 may include, but is not limited to, a modem or network card and the communication channel 190 may be implemented, for example, within a wired and / or a wireless medium.

[0042] Data is streamed to the system 100, in various embodiments, using a Wi-Fi network such as IEEE 802. 11. The Wi-Fi signal of these embodiments is received over the communications channel 190 and the communications interface 150 which are adapted for Wi-Fi communications. The communications channel 190 of these embodiments is typically connected to an access point or router that provides access to outside networks including the Internet for allowing streaming applications and other over-the-top communications. Other embodiments provide streamed data to the system 100 using a set-top box that delivers the data over the HDMI connection of the input block 105. Still other embodiments provide streamed data to the system 100 using the RF connection of the input block 105.

[0043] The system 100 may provide an output signal to various output devices, including a display 165, speakers 175, and other peripheral devices 185. The other peripheral devices 185 include, in various examples of embodiments, one or more of a stand-alone DVR, a disk player, a stereo system, a lighting system, and other devices that provide a function based on the output of the system 100. In various embodiments, control signals are communicated between the system 100 and the display 165, speakers 175, or other peripheral devices 185 using signaling such as AV. Link, CEC, or other communications protocols that enable device-to-device control with or without user intervention. The output devices may be communicatively coupled to system 100 via dedicated connections through respective interfaces 160, 170, and 180. Alternatively, the output devices may be connected to system 100 using the communications channel 190 via the communications interface 150. The display 165 and speakers 175 may be integrated in a single unit with the other components of system 100 in an electronic device, for example, a television. In various embodiments, the display interface 160 includes a display driver, for example, a timing controller (T Con) chip.

[0044] The display 165 and speaker 175 may alternatively be separate from one or more of the other components, for example, if the RF portion of input 105 is part of a separate set-top box. In various embodiments in which the display 165 and speakers 175 are external components, the output signal may be provided via dedicated output connections, including, for example, HDMI ports, USB ports, or COMP outputs.

[0045] FIG. 2 illustrates an example of a block-based hybrid video encoder 200, such as a VVC (Versatile Video Coding) encoder. FIG. 2 may also illustrate an encoder in which improvements are made to the VVC standard or an encoder employing technologies similar to VVC.

[0046] In the present application, the terms “reconstructed” and “decoded” may be used interchangeably, the terms “encoded” or “coded” may be used interchangeably, and the terms “image,” “picture” and “frame” may be used interchangeably. Usually, but not necessarily, the term “reconstructed” is used at the encoder side while “decoded” is used at the decoder side.

[0047] Before being encoded, the video sequence may go through pre-encoding processing (201), for example, applying a color transform to the input color picture (e.g., conversion from RGB 4:4:4 to YCbCr 4:2:0), or performing a remapping of the input picture components in order to get a signal distribution more resilient to compression (for instance using a histogram equalization ofone of the color components). Metadata can be associated with the pre-processing, and attached to the bitstream.

[0048] In the encoder 200, a picture is encoded by the encoder elements as described below. The picture to be encoded is partitioned (202) and processed in units of, for example, CUs (Coding Units). Each unit is encoded using, for example, either an intra or inter mode. When a unit is encoded in an intra mode, it performs intra prediction (260). In an inter mode, motion estimation (275) and compensation (270) are performed. The encoder decides (205) which one of the intra mode or inter mode to use for encoding the unit, and indicates the intra / inter decision by, for example, a prediction mode flag. Prediction residuals are calculated, for example, by subtracting (210) the predicted block from the original image block.

[0049] The prediction residuals are then transformed (225) and quantized (230). The quantized transform coefficients, as well as motion vectors and other syntax elements such as the picture partitioning information, are entropy coded (245) to output a bitstream. As a non-limiting example, context-based adaptive binary arithmetic coding (CAB AC) can be used to encode syntax elements into the bitstream.

[0050] The encoder can skip the transform and apply quantization directly to the non-transformed residual signal. The encoder can bypass both transform and quantization, i.e., the residual is coded directly without the application of the transform or quantization processes.

[0051] The encoder decodes an encoded block to provide a reference for further predictions. The quantized transform coefficients are de-quantized (240) and inverse transformed (250) to decode prediction residuals. Combining (255) the decoded prediction residuals and the predicted block, an image block is reconstructed. In-loop filters (265) are applied to the reconstructed picture to perform, for example, deblocking / SAO (Sample Adaptive Offset) / ALF (Adaptive Loop Filter) filtering to reduce encoding artifacts. The filtered image is stored in a reference picture buffer (280).

[0052] FIG. 3 illustrates a block diagram of an example video decoder 300. In the decoder 300, a bitstream is decoded by the decoder elements as described below. Video decoder 300 generally performs a decoding pass reciprocal to the encoding pass as described in FIG. 2. The encoder 200 also generally performs video decoding as part of encoding video data.

[0053] In particular, the input of the decoder includes a video bitstream, which can be generated by video encoder 200. The bitstream is first entropy decoded (330) to obtain transform coefficients, prediction modes, motion vectors, and other coded information. The picture partition information indicates how the picture is partitioned. The decoder may therefore divide (335) the picture according to the decoded picture partitioning information. The transform coefficients are dequantized (340) and inverse transformed (350) to decode the prediction residuals. Combining (355) the decoded prediction residuals and the predicted block, an image block is reconstructed. The predicted block can be obtained (370) from intra prediction (360) or motion-compensated prediction (i.e., inter prediction) (375). In-loop filters (365) are applied to the reconstructed image. The filtered image is stored at a reference picture buffer (380). Note that, for a given picture, the contents of the reference picture buffer 380 on the decoder 300 side is identical to the contents of the reference picture buffer 280 on the encoder 200 side for the same picture.

[0054] The decoded picture can further go through post-decoding processing (385), for example, an inverse color transform (e.g., conversion from YCbCr 4:2:0 to RGB 4:4:4) or an inverse remapping performing the inverse of the remapping process performed in the pre-encoding processing (201). The post-decoding processing can use metadata derived in the pre-encoding processing and signaled in the bitstream.

[0055] Intra Prediction in ECM 10 (Enhanced Compression Model 10)

[0056] To capture the arbitrary edge directions presented in natural video, ECM features 65 directional intra prediction modes as in VVC. For predicting blocks with smoothly varying textures, VVC uses the PLANAR and DC modes. These 67 core intra prediction modes as shown in FIG. 4 apply for all block sizes and in both luma and chroma intra predictions.

[0057] Intra template matching prediction (IntraTMP) is an intra prediction mode that copies a prediction block from the reconstructed part of the current frame, whose L-shaped template most closely matches the current template. For a predefined search range, the encoder searches for the most similar template to the current template in a reconstructed part of the current frame and uses the corresponding block as a prediction block. The encoder then signals the usage of this mode, and the same prediction operation is performed at the decoder side.

[0058] When DIMD (Decoder Side Intra Mode Derivation) is applied, up to five intra modes arederived from the reconstructed neighbor samples, and those five predictors are combined with the planar mode predictor with weights derived from the histogram of gradients.

[0059] In Intra Block Copy (IBC), the prediction is obtained from a reference block within the same picture indicated by a block vector.

[0060] In ECM, a CU is used for prediction and transform processing without any further partitioning except in some special cases. For example, when Intra sub-partition (ISP) is enabled in ECM, an intra CU can contain multiple TUs (Transform Units). The ISP tool divides luma intra-predicted blocks vertically or horizontally into 2 or 4 sub-partitions depending on the block size dimensions, as shown in FIG. 5 and FIG. 6, respectively. All sub-partitions fulfill the condition of having at least 16 samples. For each of these sub-partitions, a residual signal is generated by entropy decoding the coefficients sent by the encoder and then inverse quantizing and inverse transforming them. Then, the sub-partition is intra predicted and finally the corresponding reconstructed samples are obtained by adding the residual signal to the prediction signal. Therefore, the reconstructed values of each sub-partition will be available to generate the prediction of the next one, which will repeat the process and so on. All sub-partitions of the same CU share the same intra mode.

[0061] Explicit MTS in ECM

[0062] In the VVC design, DST7 and DCT8 transform kernels are used in addition to DCT2. These kernels are used for intra and inter coding. Transform skip may also be used. In particular, in addition to the conventional type-2 DCT (DCT-2), alternate transform types, type-7 DST (DST7) and type-8 DCT (DCT8), are employed in VVC. In VVC, the primary transforms are specified as separable transforms. Five different transform pairs are supported, including (DCT2, DCT2) and four MTS (Multiple Transform Selection) mode combinations, i.e., (DST7, DST7), (DST7, DCT8), (DCT8, DST7) and (DCT8, DCT8). Note the “transform pair” refers to a pair of horizontal transform and vertical transform. The combination between DCT2 and DST7 (or DCT8) is not supported in VVC.

[0063] In VVC, there are two variants of MTS, called explicit MTS and implicit MTS. The explicit MTS can be applied to both intra and inter coded blocks, while the implicit MTS can be only used for intra coded blocks. In explicit MTS, the choice of DST-7 / DCT-8 is indicated byexplicit signaling of the transform type. In implicit MTS, the transform type is selected based on coded information that is known to both the encoder and decoder, and transform type signalling is not needed.

[0064] In ECM 10, additional primary transforms including DCT5, DST4, DST1, and identity transform (IDT) are employed. Also MTS sets are made dependent on the TU size and intra mode information. For blocks predicted using the IntraTMP mode, the DIMD process is used on the prediction block to derive an intra mode that is used for transform selection. Specifically, a horizontal gradient and a vertical gradient are calculated for each predicted sample to build a Histogram of Gradients (HoG). Then the intra prediction mode with the largest histogram amplitude values is used to determine the MTS transform pair set (a set of possible transform pairs).

[0065] The ECM explicit intra MTS LUT considers 16 different TU sizes, and for each TU size, 5 different classes are considered depending on intra mode information. For each class, 1, 4 or 6 different transform pairs are considered. The number of intra MTS candidates is adaptively selected (between 1, 4 and 6 MTS candidates) depending on the sum of absolute value of transform coefficients. The sum is compared against the two fixed thresholds to determine the total number of allowed MTS candidates:(1 , if sum < thO total number of allowed MTS candidates = 4 , if th 0 < sum < thl(6 , if sum > thl

[0066] The transform pair index in the set is sent to the decoder using one, two or three bits depending on if the number of transform candidates is 1, 4 or 6, respectively.

[0067] Although a total of 80 (16 transform sizes times 5 classes) different classes are considered, some of these classes may share a same transform pair set. In ECM, there are 58 (less than 80) unique entries in the resultant look-up table (LUT).

[0068] For angular intra modes, a joint symmetry over the TU shape and intra prediction mode is considered. Particularly, a mode i (i > 34) with TU shape A*B will be mapped to the same class corresponding to the mode j = (68 - i) with TU shape B* A. However, for each transform pair the order of the horizontal and vertical transform kernel is swapped. For example, for a 16x4 block with mode 18 (horizontal prediction) and a 4x16 block with mode 50 (vertical prediction) are mapped to the same class. However, the vertical and horizontal transform kernels are swapped.For the wide-angle modes the nearest conventional angular mode is used for the transform set determination. For example, mode 2 is used for all the modes between -2 and -14. Similarly, mode 66 is used for mode 67 to mode 80. In addition, a dedicated mode index is assigned to MIP (Matrix-based Intra-picture Prediction), resulting in a total of 36 possible modes entries in the LUT. An example of transform set selection in the LUT for a given TU size and intra mode (modeldx = 4, sizeldx = 0) is illustraded in FIG. 7.

[0069] IDT may be applied for blocks that are 16x16 or smaller and have intra modes within the proximity of horizontal and vertical intra directions, where the proximity is defined by a threshold that depends on the block size. If the transform index is equal to 3 and the block satisfies the above condition, the horizontal and / or vertical identity transform is applied.

[0070] Implicit MTS in ECM

[0071] Implicit MTS is a tool in VVC that is a simpler version of the explicit MTS where a single pair of horizontal and vertical transforms is considered instead of multiple transform pairs. The benefits of the implicit MTS mode are its rapidity in an encoder compared to the explicit MTS mode, and its high compression gain when compared with the case where MTS is switched off.

[0072] In ECM, Implicit MTS is implemented by only considering DCT2 and DTS7 transform as in VVC. Specifically, if the block height or width is between 4 and 16, the vertical or horizontal transform is set to DST7, respectively. Otherwise DCT2 is selected. trTypeHor = ( nTbW > 4 && nTbW < 16 ) 7 DST7 : DCT2 trTypeVer = ( nTbH > 4 && nTbH < 16 ) 7 DST7 : DCT2

[0073] Mode-wise LUT for implicit MTS

[0074] As described above, in ECM, four new primary transforms, namely DCT5, DST4, DST1, and IDT, are added to MTS, which was limited to DCT8 and DST7 in VVC. Although bringing a rate-distortion gain, introducing new transforms increases the encoder complexity as additional tests may be required during the rate-distortion optimization loop. To overcome this, implicit MTS can be employed where the best transform is inferred based on information known on the decoder side. However, introducing new transforms increases the solution space and makes implicit MTS more challenging compared to explicit MTS, where the best transform pair is signaled to the decoder. To improve the performance of implicit MTS, we propose an implicit MTS method basedon a mode-wise lookup table (LUT). This LUT infers the best transform pair for a given block from its intra prediction mode and TU size. The LUT is trained offline using statistics collected, e.g., from an exhaustive search on transforms performed in the encoder. This approach offers a flexible distribution of transform pairs between consecutive modes and allows combinations of other transforms with DCT2, which are not considered in the current ECM explicit MTS LUT. The LUT design method is applied to both the encoder and decoder.

[0075] FIG. 8 illustrates the decoder-side process of the proposed implicit MTS using mode-wise LUT according to this embodiment. Unlike in ECM where explicit intra MTS uses a LUT to assign a transform pair set to a block based on its intra prediction mode and TU size, the proposed mode-wise LUT assigns a single transform pair per block based on its intra prediction mode and TU size. This way, by assigning different transform pairs for consecutive intra modes, the proposed LUT introduces diversity in the search space while avoiding the need for a transform selection process. Intra prediction modes may include planar, DC, MIP, angular modes, and wide angles. A fixed range of TU sizes is considered (e.g., 4 < width < 32 and 4 < height < 32).

[0076] In particular, transformed coefficients are obtained (810) and the transform pair is inferred (820) from the LUT using the intra prediction mode and the TU size of the residual block. Then, inverse transform is performed (830) on the transform coded block using the corresponding transform pair to reconstruct (840) the residual block.

[0077] On the encoder side, the transform pair is inferred from the LUT using the intra prediction mode and the TU size of the residual block. Then, the corresponding transform pair is applied to the residual block to produce the transformed residual block. Since the same intra prediction mode and TU size are known at the encoder and decoder, the same transform pair can be determined based solely on the intra prediction mode and the TU size in the same manner at the encoder and decoder, and no signaling overhead is required for conveying the transform pair information from the encoder to decoder.

[0078] In a variant, only the TU size is used to derive the transform pair in the LUT. Alternatively, only the intra prediction mode is used to derive the transform pair in the LUT.

[0079] FIG. 9 illustrates a hybrid explicit / implicit method on the encoder side, according to an embodiment. After the residual block is obtained (910), the encoder checks whether explicit MTSis selected (920). If explicit MTS is enabled for the intra prediction mode of the residual block, intra explicit MTS is performed (940) and the index of the selected transform pair (mts idx) is signaled in the bitstream. If transform skip is used (930), no transform operation is performed on the residual signal and a transform skip flag (ts flag) is sent in the bitstream. If implicit MTS is used, the transform pair is inferred (970) from the LUT using the intra prediction mode and the TU size of the residual block and no index or flag is signaled in the bitstream. The residuals are transformed (950) with the selected transform pair and the transform coefficients are quantized (960) before being entropy encoded.

[0080] FIG. 10 illustrates a hybrid explicit / implicit method on the decoder side, according to an embodiment. In particular, transformed coefficients are obtained (1010). If explicit MTS is used (1020, mts idx > 1), the MTS transform pair is decoded (1040) from explicit signaling. If transform skip is used (mts idx is set to 1 if ts flag is true, 1030), no transform pair is decoded nor inferred, and inverse transform (1050) is skipped. If implicit MTS is used (mts idx = 0), the transform pair is inferred (1070) from the LUT using the intra prediction mode and the TU size of the transform coded block. Then, inverse transform is performed (1050) on the transform coded block using the corresponding transform pair to reconstruct (1060) the residual block.

[0081] The process of extracting the syntax elements from the bitstream is called the parsing process. The parsing may just be limited to the decoding of syntax elements values, where bits from the bitstream are used as inputs, syntax element values are provided as outputs. For each element, a descriptor is used in the syntax table to specify the applicable parsing process.

[0082] For example, in VVC, the following descriptors specify the parsing process of each syntax element:- ae(v): context-adaptive arithmetic entropy-coded syntax element.- b(8): byte having any pattern of bit string (8 bits).- f(n): fixed-pattern bit string using n bits written (from left to right) with the left bit first.- i(n) : signed integer using n bits. When n is “v” in the syntax table, the number of bits varies in a manner dependent on the value of other syntax elements.- se(v): signed integer O-th order Exp-Golomb-coded syntax element with the left bit first.- u(n): unsigned integer using n bits. When n is “v” in the syntax table, the number of bits varies in a manner dependent on the value of other syntax elements.- ue(v): unsigned integer O-th order Exp-Golomb-coded syntax element with the left bit first.

[0083] The decoding process specifies how the syntax elements are used to reconstruct the samples. The decoding process takes the syntax elements in the bitstream as input, and reconstructs the video sequence based on the semantics of the syntax elements.

[0084] When designing a video codec, one rule is to achieve independent parsing, where the parsing process is independent of the decoding process. Generally, the partitioning between parsing and decoding processes is governed by the rule of limiting resources for parsing in order to dedicate lightweight hardware or software resources to the parsing process.

[0085] In one example, some modes known at the parsing stage (e.g., planar and / or MIP) use explicit MTS while the other modes use the proposed mode-wise implicit MTS LUT. It should be noted all or a subset of the modes that are known at the parsing stage will use explicit MTS. If a subset of the modes from the parsing stages uses explicit MTS, then the remaining modes available from parsing will use implicit MTS or DCT2 (e.g., implicit MTS is disabled for MIP in ECM). In case a mode uses explicit MTS, a LUT for intra explicit MTS is used. An MTS index may be parsed for each CU or block that uses a mode known at the parsing stage if the known mode belongs to the subset of modes that use explicit MTS. No such MTS index is parsed for the other blocks.

[0086] In a variant, the proposed method is only applied to ISP blocks.

[0087] In the ECM, DIMD and TIMD (Template-based Intra Mode Derivation) can be used to derive an intra prediction mode from already coded samples. The original purpose of these tools is to offer the option to avoid signaling the intra mode to the decoder. However, they can also be used to obtain an intra mode index in scenarios where it is not available, for instance, IntraTMP, IBC, and inter blocks. Therefore, those modes can use tools that rely on a LUT with intra-mode as input, like intra-explicit MTS or LFNST. The mode obtained using this approach is denoted as a virtual intra mode. As a variant, the proposed method can be applied to IntraTMP, IBC, and inter blocks using a virtual intra prediction mode.

[0088] Designing a LUT based on offline transform analysis

[0089] The ECM uses a LUT for intra explicit MTS containing transform sets shared across fixed ranges of intra modes. However, the ECM explicit MTS LUT does not include any transform pairthat combines DCT2 with a non-DCT2 transform. However, such combinations might be suitable when the residual block is obtained from a horizontal or vertical intra-prediction. Generally, residual samples have larger absolute values if they are farther away from the reference samples. Therefore, an increasing transform along the intra prediction direction (such as DST7) and a constant transform along the opposite intra prediction direction (such as DCT2) might better catch the residual statistics.

[0090] To address this issue, we design a new LUT dedicated to intra implicit MTS. To create the mode-wise LUT based on transform pairs statistics, an offline search is performed using intra explicit MTS on a fixed video sequence dataset. The best transform pair per size / mode (i.e., the transform pair with the highest number of occurrences during the search) is selected as the candidate in the LUT. No index is sent during the search to avoid signaling bias, and only encoder statistics are considered.

[0091] In a variant, an index is slowly introduced during the learning process and the rate at which the index is introduced is based on a parameter (also known as annealing rate). This allows the encoder to start learning without any bias to allow more randomness and slowly reduce the randomness. This method is also known as simulated annealing and is used for solving optimization problems.

[0092] In a variant, an exhaustive search based on all possible transform combinations is performed to remove the bias from the current explicit MTS LUT. It also enables combinations of non-DCT2 and DCT2 transforms.

[0093] In a variant, other trigonometric transforms (i.e., other than DCT5, DST4, DST1, DCT8, DST7, and DCT2) are introduced during the exhaustive search and can be used as candidate in the LUT.

[0094] FIG. 11 illustrates the training process for one block, according to an embodiment. In this embodiment, the candidate in the LUT is selected based on the rate-distortion (RD) cost of transforms during the search. For each block of the training dataset, we collect the RD cost and record it in the corresponding intra prediction mode and TU size bin in the LUT (1180). In particular, the intra predicted block is obtained (1110) and the residual block is obtained (1120) for the block. Tmax corresponds to the last transform pair to be tested during the exhaustive search.For any transform pair Ti to be tested (1130), the transform coefficient using the transform pair Ti is obtained (1140) and quantized (1145). De-quantization (1155) and inverse transform (1160) are applied. The block is reconstructed (1160) and the RD cost is computed (1170). The processing of one block ends (1199) after the RD cost is saved in the corresponding intra prediction mode and TU size bin in the LUT (1180). At the end (1199) of the training process (after all blocks from the training dataset have been processed), the pair with the lowest RD cost at each LUT bin is selected as the candidate in the mode-wise LUT. It allows collecting statistics on transform pairs that are not selected during the search. An example of transform pair ranking based on the RD-cost is given in FIG. 12.

[0095] In a variant, a specific LUT can be trained for the hybrid explicit / implicit use case. Here, the exhaustive search is performed only on modes that are considered implicit modes (e.g., non- planar and / or non-MIP modes). In this context, planar and MIP modes use the default intra explicit MTS during the search. It allows matching the behavior of these modes when hybrid explicit / implicit MTS is performed.

[0096] In another variant, the statistics of ISP modes are not considered in the construction of the LUT since they have their own implicit MTS method.

[0097] Using post-processing to enhance the LUT

[0098] This embodiment proposes to apply post-processing to the collected statistics to enhance the LUT. One way to achieve this is by applying noise to the statistics. This noise can be applied to either the number of transform occurrences or the RD cost statistics, depending on the method used to collect the statistics. For instance, the noise can be a Gaussian noise configured with a mean p=0 and a variance c that depends on the range of the statistics. For instance, if the statistics are considering RD cost, the variance can be set to c = 10'9. In a variant, the variance can be determined by the difference between the statistics of the first transform and the N ranked transform.

[0099] This post-processing step increases the diversity among the tested modes, which means that the transforms assigned to consecutive intra modes are more different. It can be beneficial when testing consecutive intra modes as it gives more transform options to the encoder with limited impact on the performance. In a variant, the noise is applied only to a subset of the rankedtransform pairs, such as the first N best transform pairs.

[0100] In another variant, the second-best transform is selected if the number of occurrences of the best one is lower than the second-best one divided by a certain threshold.

[0101] FIG. 13 A and FIG. 13B show examples of LUTs represented as heatmaps without and with noise added, respectively, where each bin represents a transform pair for a given intra mode and TU size. These figures demonstrate that applying noise to the statistics allows for higher diversity of transform pairs among consecutive intra modes. A higher diversity is observed because consecutive intra modes have more different transform assigned in FIG. 13B than in FIG. 13 A.

[0102] Coefficients of the LUT

[0103] The transform pair indices mapped to each intra mode and TU size combinations have been obtained by using the training process described before. The training is based on an exhaustive analysis of transform in ECM-8.0 using RD-cost. No post-processing has been applied to obtain these coefficients. The transform pair index for each intra mode and TU size are represented in the tables in FIGs. 14A-14D. The corresponding transform pairs are represented in Table 1.

[0104] The performance is assessed on top of the enhanced compression model (ECM- 10.0) software. To allow a fair comparison with state-of-the-art approaches based on VTM (VVC Test Model), we also compare our method with a modified version of the ECM- 10.0. This configuration, denoted as ECM*, limits MTS to transform pairs composed of DST7 and DCT8, as in the last version of the VTM. To evaluate the performance of our approach, we use the Bjontegaard-Delta rate (BD-rate) method based on the peak signal-to-noise ratio (PSNR) computed on the luma component. All the results are given following the common test conditions (CTCs) using all intra (Al) configurations. FIG. 15 displays the BD-rate / complexity tradeoff of each approach in a 2D plane.Table 1

[0105] Regarding hybrid configurations, the proposed LUT -based method outperforms the default ECM implicit MTS for all hybrid configurations. Moreover, all the proposed LUT -based hybrid approaches have lower encoding complexity than their corresponding default implicit MTS versions. This is because the LUT is more efficient than the default implicit MTS ECM method, making implicit MTS selected more often.

[0106] Regarding FIG. 15, enabling explicit MTS only for MIP and using the proposed LUT for all the other modes (Hyb. MIP) appears to be the best tradeoff among the hybrid scenarios. This approach offers a better RD performance than the planar hybrid method (Hyb. PL), given that the MIP mode uses DCT2 / DCT2 in the planar hybrid approach. Finally, it is more attractive thanenabling explicit MTS for both planar and MIP (Hyb. PL / MIP), as the individual gains of both MIP and planar hybrid methods are not additive.

[0107] The results show that the proposed hybrid approach reduces the ECM complexity by 16% with a luma BD-rate loss of 0.40% and 0.15% compared to the MTS method of ECM and VTM, respectively.

[0108] Various methods are described herein, and each of the methods comprises one or more steps or actions for achieving the described method. Unless a specific order of steps or actions is required for proper operation of the method, the order and / or use of specific steps and / or actions may be modified or combined. Additionally, terms such as “first”, “second”, etc. may be used in various embodiments to modify an element, component, step, operation, etc., for example, a “first decoding” and a “second decoding”. Use of such terms does not imply an ordering to the modified operations unless specifically required. So, in this example, the first decoding need not be performed before the second decoding, and may occur, for example, before, during, or in an overlapping time period with the second decoding.

[0109] Various methods and other aspects described in this application can be used to modify modules, for example, the transform and inverse transform modules (225, 250, 350), of a video encoder 200 and decoder 300 as shown in FIG. 2 and FIG. 3. Moreover, the present aspects are not limited to ECM, VVC or HEVC, and can be applied, for example, to other standards and recommendations, and extensions of any such standards and recommendations. Unless indicated otherwise, or technically precluded, the aspects described in this application can be used individually or in combination.[HO] Various numeric values are used in the present application. The specific values are for example purposes and the aspects described are not limited to these specific values.[Hl] Various implementations involve decoding. “Decoding,” as used in this application, may encompass all or part of the processes performed, for example, on a received encoded sequence in order to produce a final output suitable for display. In various embodiments, such processes include one or more of the processes typically performed by a decoder, for example, entropy decoding, inverse quantization, inverse transformation, and differential decoding. Whether the phrase “decoding process” is intended to refer specifically to a subset of operations or generally tothe broader decoding process will be clear based on the context of the specific descriptions and is believed to be well understood by those skilled in the art.

[0112] Various implementations involve encoding. In an analogous way to the above discussion about “decoding”, “encoding” as used in this application may encompass all or part of the processes performed, for example, on an input video sequence in order to produce an encoded bitstream.

[0113] The implementations and aspects described herein may be implemented in, for example, a method or a process, an apparatus, a software program, a data stream, or a signal. Even if only discussed in the context of a single form of implementation (for example, discussed only as a method), the implementation of features discussed may also be implemented in other forms (for example, an apparatus or program). An apparatus may be implemented in, for example, appropriate hardware, software, and firmware. The methods may be implemented in, for example, an apparatus, for example, a processor, which refers to processing devices in general, including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device. Processors also include communication devices, for example, computers, cell phones, portable / personal digital assistants (“PDAs”), and other devices that facilitate communication of information between end-users.

[0114] Reference to “one embodiment” or “an embodiment” or “one implementation” or “an implementation”, as well as other variations thereof, means that a particular feature, structure, characteristic, and so forth described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of the phrase “in one embodiment” or “in an embodiment” or “in one implementation” or “in an implementation”, as well any other variations, appearing in various places throughout this application are not necessarily all referring to the same embodiment.

[0115] Additionally, this application may refer to “determining” various pieces of information. Determining the information may include one or more of, for example, estimating the information, calculating the information, predicting the information, or retrieving the information from memory.

[0116] Further, this application may refer to “accessing” various pieces of information. Accessing the information may include one or more of, for example, receiving the information, retrieving the information (for example, from memory), storing the information, moving the information,copying the information, calculating the information, determining the information, predicting the information, or estimating the information.

[0117] Additionally, this application may refer to “receiving” various pieces of information. Receiving is, as with “accessing”, intended to be a broad term. Receiving the information may include one or more of, for example, accessing the information, or retrieving the information (for example, from memory). Further, “receiving” is typically involved, in one way or another, during operations, for example, storing the information, processing the information, transmitting the information, moving the information, copying the information, erasing the information, calculating the information, determining the information, predicting the information, or estimating the information.

[0118] It is to be appreciated that the use of any of the following “ / ”, “and / or”, and “at least one of’, for example, in the cases of “A / B”, “A and / or B” and “at least one of A and B”, is intended to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of both options (A and B). As a further example, in the cases of “A, B, and / or C” and “at least one of A, B, and C”, such phrasing is intended to encompass the selection of the first listed option (A) only, or the selection of the second listed option (B) only, or the selection of the third listed option (C) only, or the selection of the first and the second listed options (A and B) only, or the selection of the first and third listed options (A and C) only, or the selection of the second and third listed options (B and C) only, or the selection of all three options (A and B and C). This may be extended, as is clear to one of ordinary skill in this and related arts, for as many items as are listed.

[0119] Also, as used herein, the word “signal” refers to, among other things, indicating something to a corresponding decoder. For example, in certain embodiments the encoder signals a quantization matrix for de-quantization. In this way, in an embodiment the same parameter is used at both the encoder side and the decoder side. Thus, for example, an encoder can transmit (explicit signaling) a particular parameter to the decoder so that the decoder can use the same particular parameter. Conversely, if the decoder already has the particular parameter as well as others, then signaling can be used without transmitting (implicit signaling) to simply allow the decoder to know and select the particular parameter. By avoiding transmission of any actual functions, a bit savings is realized in various embodiments. It is to be appreciated that signaling can be accomplished ina variety of ways. For example, one or more syntax elements, flags, and so forth are used to signal information to a corresponding decoder in various embodiments. While the preceding relates to the verb form of the word “signal”, the word “signal” can also be used herein as a noun.

[0120] As will be evident to one of ordinary skill in the art, implementations may produce a variety of signals formatted to carry information that may be, for example, stored or transmitted. The information may include, for example, instructions for performing a method, or data produced by one of the described implementations. For example, a signal may be formatted to carry the bitstream of a described embodiment. Such a signal may be formatted, for example, as an electromagnetic wave (for example, using a radio frequency portion of spectrum) or as a baseband signal. The formatting may include, for example, encoding a data stream and modulating a carrier with the encoded data stream. The information that the signal carries may be, for example, analog or digital information. The signal may be transmitted over a variety of different wired or wireless links, as is known. The signal may be stored on a processor-readable medium.

Claims

CLAIMS1. A method of video decoding, comprising: selecting a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; inverse transforming transform coefficients of said current block based on said single transform pair; obtaining residuals for said current block from said inverse transformed coefficients; and decoding said current block based on said residuals.

2. A method of video encoding, comprising: selecting a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; obtaining residuals for said current block; transforming said residuals for said current block to obtain transform coefficients of said current block based on said single transform pair; and encoding said current block based on said transform coefficients.

3. An apparatus for video decoding, comprising at least a memory and one or more processors, wherein said one or more processors are configured to: select a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes;inverse transform transform coefficients of said current block based on said single transform pair; obtain residuals for said current block from said inverse transformed coefficients; and decode said current block based on said residuals.

4. An apparatus for video encoding, comprising at least a memory and one or more processors, wherein said one or more processors are configured to: select a single transform pair from a plurality of transform pairs for a current block, wherein said selecting is only based on at least one of an intra prediction mode and a transform unit size associated with said current block, and said intra prediction mode is selected from a plurality of intra prediction modes and said transform unit size is selected from a plurality of transform unit sizes; obtain residuals for said current block; transform said residuals for said current block to obtain transform coefficients of said current block based on said single transform pair; and encode said current block based on said transform coefficients.

5. The method of claim 1 or 2, or the apparatus of claim 3 or 4, wherein said single transform pair is selected based on a look up table, wherein said look up table indicates a mapping between (1) said at least one of said intra prediction mode and said transform unit size and (2) said single transform pair.

6. The method of claim 1, 2 or 5, or the apparatus of any one of claims 3-5, wherein said plurality of transform pairs include a pair of transforms, which includes a DCT2 transform and another trigonometric transform.

7. The method of any one of claims 1, 2, 5 and 6, or the apparatus of any one of claims 3-6, wherein said plurality of intra prediction modes include all intra prediction modes.

8. The method of any one of claims 1, 2 and 5-6, or the apparatus of any one of claims 3-6, wherein a subset of intra prediction modes is available at a parsing stage, and wherein said plurality of intra prediction modes only include intra prediction modes excluding said subset of intra prediction modes.

9. The method of claim 8, or the apparatus of claim 8, wherein said subset of intra prediction modes corresponds to at least one of PLANAR and MIP modes.

10. The method of claim 8, or the apparatus of claim 8, wherein said plurality of intra prediction modes include at least one intra prediction mode that is available at parsing.

11. The method of any one of claims 8-10, or the apparatus of any one of claims 8-10, wherein explicit signaling is used for a transform pair corresponding to a TU with an intra prediction mode of said subset of intra prediction modes.

12. The method of any one of claims 1, 2 and 5-11, or the apparatus of any one of claims 3-11, wherein said plurality of transform unit sizes include all possible TU sizes.

13. A signal comprising video data, formed by performing the method of any one of claims 2 and 5-12.

14. A computer readable storage medium having stored thereon instructions for encoding or decoding a video according to the method of any one of claims 1, 2 and 5-12.

Images