Method and apparatus of latency reduction for cross-component model and residual scaling for in-loop filtering in video coding
Parallel execution of ALF and CCM processes, combined with residual scaling, addresses latency issues in video coding systems, ensuring efficient and timely decoding.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- MEDIATEK INC
- Filing Date
- 2025-11-04
- Publication Date
- 2026-05-07
AI Technical Summary
Existing video coding systems utilizing Adaptive Loop Filter (ALF) and Cross-Component Model (CCM) face processing latency issues, particularly at the decoder, which can be exacerbated by additional stages introduced in methods like cross-component convolutional models and residual scaling.
Perform in-loop filtering processes, such as ALF and CCM, in parallel, and apply residual scaling to reduce latency by using weighted sums or separate scaling factors for ALF and CCALF processes, allowing these processes to operate independently without relying on each other's intermediate results.
Reduces processing latency by enabling simultaneous execution of ALF and CCM operations, maintaining coding benefits while minimizing decoder latency.
Smart Images

Figure CN2025132436_07052026_PF_FP_ABST
Abstract
Description
METHOD AND APPARATUS OF LATENCY REDUCTION FOR CROSS-COMPONENT MODEL AND RESIDUAL SCALING FOR IN-LOOP FILTERING IN VIDEO CODINGCROSS REFERENCE TO RELATED APPLICATIONS
[0001] The present invention is a non-Provisional Application of and claims priority to U.S. Provisional Patent Application No. 63 / 715,676, filed on November 4, 2024. The U.S. Provisional Patent Application is hereby incorporated by reference in its entirety.FIELD OF THE INVENTION
[0002] The present invention relates to video coding system using ALF (Adaptive Loop Filter) . In particular, the present invention discloses schemes to reduce processing latency for video coding systems utilising ALF and CCM (Cross-Component Model) . BACKGROUND AND RELATED ART
[0003] Versatile video coding (VVC) is the latest international video coding standard developed by the Joint Video Experts Team (JVET) of the ITU-T Video Coding Experts Group (VCEG) and the ISO / IEC Moving Picture Experts Group (MPEG) . The standard has been published as an ISO standard: ISO / IEC 23090-3: 2021, Information technology -Coded representation of immersive media -Part 3: Versatile video coding, published Feb. 2021. VVC is developed based on its predecessor HEVC (High Efficiency Video Coding) by adding more coding tools to improve coding efficiency and also to handle various types of video sources including 3-dimensional (3D) video signals.
[0004] Fig. 1A illustrates an exemplary adaptive Inter / Intra video encoding system incorporating loop processing. For Intra Prediction, the prediction data is derived based on previously coded video data in the current picture. For Inter Prediction 112, Motion Estimation (ME) is performed at the encoder side and Motion Compensation (MC) is performed based on the result of ME to provide prediction data derived from other picture (s) and motion data. Switch 114 selects Intra Prediction 110 or Inter-Prediction 112 and the selected prediction data is supplied to Adder 116 to form prediction errors, also called residues. The prediction error is then processed by Transform (T) 118 followed by Quantization (Q) 120. The transformed and quantized residues are then coded by Entropy Encoder 122 to be included in a video bitstream corresponding to the compressed video data. The bitstream associated with the transform coefficients is then packed with side information such as motion and coding modes associated with Intra prediction and Inter prediction, and other information such as parameters associated with loop filters applied to underlying image area. The side information associated with Intra Prediction 110, Inter prediction 112 and in-loop filter 130, are provided to Entropy Encoder 122 as shown in Fig. 1A. When an Inter-prediction mode is used, a reference picture or pictures have to be reconstructed at the encoder end as well. Consequently, the transformed and quantized residues are processed by Inverse Quantization (IQ) 124 and Inverse Transformation (IT) 126 to recover the residues. The residues are then added back to prediction data 136 at Reconstruction (REC) 128 to reconstruct video data. The reconstructed video data may be stored in Reference Picture Buffer 134 and used for prediction of other frames.
[0005] As shown in Fig. 1A, incoming video data undergoes a series of processing in the encoding system. The reconstructed video data from REC 128 may be subject to various impairments due to a series of processing. Accordingly, in-loop filter 130 is often applied to the reconstructed video data before the reconstructed video data are stored in the Reference Picture Buffer 134 in order to improve video quality. For example, deblocking filter (DF) , Sample Adaptive Offset (SAO) and Adaptive Loop Filter (ALF) may be used. The loop filter information may need to be incorporated in the bitstream so that a decoder can properly recover the required information. Therefore, loop filter information is also provided to Entropy Encoder 122 for incorporation into the bitstream. In Fig. 1A, Loop filter 130 is applied to the reconstructed video before the reconstructed samples are stored in the reference picture buffer 134. The system in Fig. 1A is intended to illustrate an exemplary structure of a typical video encoder. It may correspond to the High Efficiency Video Coding (HEVC) system, VP8, VP9, H. 264 or VVC.
[0006] The decoder, as shown in Fig. 1B, can use similar or portion of the same functional blocks as the encoder except for Transform 118 and Quantization 120 since the decoder only needs Inverse Quantization 124 and Inverse Transform 126. Instead of Entropy Encoder 122, the decoder uses an Entropy Decoder 140 to decode the video bitstream into quantized transform coefficients and needed coding information (e.g. ILPF information, Intra prediction information and Inter prediction information) . The Intra prediction 150 at the decoder side does not need to perform the mode search. Instead, the decoder only needs to generate Intra prediction according to Intra prediction information received from the Entropy Decoder 140. Furthermore, for Inter prediction, the decoder only needs to perform motion compensation (MC 152) according to Inter prediction information received from the Entropy Decoder 140 without the need for motion estimation.
[0007] According to VVC, an input picture is partitioned into non-overlapped square block regions referred as CTUs (Coding Tree Units) , similar to HEVC. Each CTU can be partitioned into one or multiple smaller size coding units (CUs) . The resulting CU partitions can be in square or rectangular shapes. Also, VVC divides a CTU into prediction units (PUs) as a unit to apply prediction process, such as Inter prediction, Intra prediction, etc.
[0008] Adaptive Loop Filter in VVC
[0009] In VVC, an Adaptive Loop Filter (ALF) with block-based filter adaption is applied. For the luma component, one filter is selected among 25 filters for each 4×4 block, based on the direction and activity of local gradients.
[0010] 1. Filter shape
[0011] Two diamond filter shapes (as shown in Fig. 2) are used. The 7×7 diamond shape 220 is applied for luma component and the 5×5 diamond shape 210 is applied for the chroma components.
[0012] 2. Block classification
[0013] For luma component, each 4×4 block is categorized into one out of 25 classes. The classification index C is derived based on its directionality D and a quantized value of activity as follows:
[0014] To calculate D and gradients of the horizontal, vertical and two diagonal direction are first calculated using 1-D Laplacian: where indices i and j refer to the coordinates of the upper left sample within the 4×4 block and R (i, j) indicates a reconstructed sample at coordinate (i, j) .
[0015] To reduce the complexity of block classification, the subsampled 1-D Laplacian calculation is applied to the vertical direction (Fig. 3A) and the horizontal direction (Fig. 3B) . As shown in Figs. 3C-D, the same subsampled positions are used for gradient calculation of all directions (gd1 in Fig. 3C and gd2 in Fig. 3D) .
[0016] Then D maximum and minimum values of the gradients of horizontal and vertical directions are set as:
[0017] The maximum and minimum values of the gradient of two diagonal directions are set as:
[0018] To derive the value of the directionality D, these values are compared against each other and with two thresholds t1 and t2: Step 1. If both and are true, D is set to 0. Step 2. If continue from Step 3; otherwise continue from Step 4. Step 3. If D is set to 2; otherwise D is set to 1. Step 4. If D is set to 4; otherwise D is set to 3.
[0019] The activity value A is calculated as:
[0020] A is further quantized to the range of 0 to 4, inclusively, and the quantized value is denoted as
[0021] For the chroma components in a picture, no classification is applied.
[0022] 3. Geometric transformations of filter coefficients and clipping values
[0023] Before filtering each 4×4 luma block, geometric transformations such as rotation or diagonal and vertical flipping are applied to the filter coefficients f (k, l) and to the corresponding filter clipping values c (k, l) depending on gradient values calculated for that block. This is equivalent to applying these transformations to the samples in the filter support region. The idea is to make different blocks to which ALF is applied more similar by aligning their directionality.
[0024] Three geometric transformations, including diagonal, vertical flip and rotation are introduced: Diagonal: fD (k, l) =f (l, k) , cD (k, l) =c (l, k) , Vertical flip: fV (k, l) =f (k, K-l-1) , cV (k, l) =c (k, K-l-1) , Rotation: fR (k, l) =f (K-l-1, k) , cR (k, l) =c (K-l-1, k) , where K is the size of the filter and 0≤k, l≤K-1 are coefficients coordinates, such that location (0, 0) is at the upper left corner and location (K-1, K-1) is at the lower right corner. The transformations are applied to the filter coefficients f (k, l) and to the clipping values c (k, l) depending on gradient values calculated for that block. The relationship between the transformation and the four gradients of the four directions are summarized in the following table. Table 1. Mapping of the gradient calculated for one block and the transformations
[0025] 4. Filtering process
[0026] At decoder side, when ALF is enabled for a CTB, each sample R (i, j) within the CU is filtered, resulting in sample value R′ (i, j) as shown below, where f (k, l) denotes the decoded filter coefficients, K (x, y) is the clipping function and c (k, l) denotes the decoded clipping parameters. The variable k and l varies between –L / 2 and L / 2, where L denotes the filter length. The clipping function K (x, y) =min (y, max (-y, x) ) which corresponds to the function Clip3 (-y, y, x) . The clipping operation introduces non-linearity to make ALF more efficient by reducing the impact of neighbour sample values that are too different with the current sample value.
[0027] 5. Cross component adaptive loop filter
[0028] CC-ALF uses luma sample values to refine each chroma component by applying an adaptive, linear filter to the luma channel and then using the output of this filtering operation for chroma refinement. Fig. 4A provides a system level diagram of the CC-ALF process with respect to the SAO, luma ALF and chroma ALF processes. As shown in Fig. 4A, each colour component (i.e., Y, Cb and Cr) is processed by its respective SAO (i.e., SAO Luma 410, SAO Cb 412 and SAO Cr 414) . After SAO, ALF Luma 420 is applied to the SAO-processed luma and ALF Chroma 430 is applied to SAO-processed Cb and Cr. However, there is a cross-component term from luma to a chroma component (i.e., CC-ALF Cb 422 and CC-ALF Cr 424) . The outputs from the cross-component ALF are added (using adders 432 and 434 respectively) to the outputs from ALF Chroma 430.
[0029] Filtering in CC-ALF is accomplished by applying a linear, diamond shaped filter (e.g. filters 440 and 442 in Fig. 4B) to the luma channel. In Fig. 4B, a blank circle indicates a luma sample and a dot-filled circle indicate a chroma sample. One filter is used for each chroma channel, and the operation is expressed as: where (x, y) is chroma component i location being refined, (xY, yY) is the luma location based on (x, y) , Si is filter support area in luma component, and ci (x0, y0) represents the filter coefficients.
[0030] As shown in Fig, 4B, the luma filter support is the region collocated with the current chroma sample after accounting for the spatial scaling factor between the luma and chroma planes.
[0031] In the VVC reference software, CC-ALF filter coefficients are computed by minimizing the mean square error of each chroma channel with respect to the original chroma content. To achieve this, the VTM (VVC Test Model) algorithm uses a coefficient derivation process similar to the one used for chroma ALF. Specifically, a correlation matrix is derived, and the coefficients are computed using a Cholesky decomposition solver in an attempt to minimize a mean square error metric. In designing the filters, a maximum of 8 CC-ALF filters can be designed and transmitted per picture. The resulting filters are then indicated for each of the two chroma channels on a CTU basis.
[0032] Additional characteristics of CC-ALF include: · The design uses a 3x4 diamond shape with 8 taps. · Seven filter coefficients are transmitted in the APS. · Each of the transmitted coefficients has a 6-bit dynamic range and is restricted to power-of-2 values. · The eighth filter coefficient is derived at the decoder such that the sum of the filter coefficients is equal to 0. · An APS may be referenced in the slice header. · CC-ALF filter selection is controlled at CTU-level for each chroma component · Boundary padding for the horizontal virtual boundaries uses the same memory access pattern as luma ALF.
[0033] As an additional feature, the reference encoder can be configured to enable some basic subjective tuning through the configuration file. When enabled, the VTM attenuates the application of CC-ALF in regions that are coded with high QP and are either near mid-grey or contain a large amount of luma high frequencies. Algorithmically, this is accomplished by disabling the application of CC-ALF in CTUs where any of the following conditions are true: · The slice QP value minus 1 is less than or equal to the base QP value. · The number of chroma samples for which the local contrast is greater than (1 << (bitDepth –2 ) ) –1 exceeds the CTU height, where the local contrast is the difference between the maximum and minimum luma sample values within the filter support region. · More than a quarter of chroma samples are in the range between (1 << (bitDepth –1) ) –16 and (1 << (bitDepth –1) ) + 16
[0034] The motivation for this functionality is to provide some assurance that CC-ALF does not amplify artefacts introduced earlier in the decoding path (This is largely due the fact that the VTM currently does not explicitly optimize for chroma subjective quality) . It is anticipated that alternative encoder implementations may either not use this functionality or incorporate alternative strategies suitable for their encoding characteristics.
[0035] 6. Filter parameters signalling
[0036] ALF filter parameters are signalled in Adaptation Parameter Set (APS) . In one APS, up to 25 sets of luma filter coefficients and clipping value indexes, and up to eight sets of chroma filter coefficients and clipping value indexes can be signalled. To reduce bits overhead, filter coefficients of different classification for luma component can be merged. In slice header, the indices of the APSs used for the current slice are signalled.
[0037] Clipping value indexes, which are decoded from the APS, allow determining clipping values using a table of clipping values for both the luma and chroma components. These clipping values are dependent of the internal bitdepth. More precisely, the clipping values are obtained by the following formula: AlfClip= {round (2B-α*n) for n∈ [0.. N-1] } with B equal to the internal bitdepth, α is a pre-defined constant value equal to 2.35, and N equal to 4 which is the number of allowed clipping values in VVC. The AlfClip is then rounded to the nearest value with the format of power of 2.
[0038] In slice header, up to 7 APS indices can be signalled to specify the luma filter sets that are used for the current slice. The filtering process can be further controlled at CTB level. A flag is always signalled to indicate whether ALF is applied to a luma CTB. A luma CTB can choose a filter set among 16 fixed filter sets and the filter sets from APSs. A filter set index is signalled for a luma CTB to indicate which filter set is applied. The 16 fixed filter sets are pre-defined and hard-coded in both the encoder and the decoder.
[0039] For the chroma component, an APS index is signalled in slice header to indicate the chroma filter sets being used for the current slice. At CTB level, a filter index is signalled for each chroma CTB if there is more than one chroma filter set in the APS.
[0040] The filter coefficients are quantized with norm equal to 128. In order to restrict the multiplication complexity, a bitstream conformance is applied so that the coefficient value of the non-central position shall be in the range of -27 to 27 -1, inclusive. The central position coefficient is not signalled in the bitstream and is considered as equal to 128.
[0041] Adaptive Loop Filter in ECM
[0042] In ECM8 (Muhammed Coban, et al., “Algorithm description of Enhanced Compression Model 8 (ECM 8) ” , Joint Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29) , 29th Meeting, by teleconference, 11–20 January 2023, Document: JVET-AC2025) , some changes from the VVC ALF are disclosed. A brief overview is shown below.
[0043] 1. ALF simplification removal
[0044] ALF gradient subsampling and ALF virtual boundary processing are removed. Block size for classification is reduced from 4x4 to 2x2. Filter size for both the luma and chroma, for which ALF coefficients are signalled, is increased to 9x9.
[0045] 2. ALF with fixed filters
[0046] To filter a luma sample, three different classifiers (C0, C1 and C2) and three different sets of filters (F0, F1 and F2) are used. Sets F0 and F1 contain fixed filters, with coefficients trained for classifiers C0 and C1. Coefficients of filters in F2 are signalled. Which filter from a set Fi is used for a given sample is decided by a class Ci assigned to this sample using classifier Ci.
[0047] 3. Filtering
[0048] At first, two 13x13 diamond shape fixed filters F0 and F1 are applied to derive two intermediate samples R0 (x, y) and R1 (x, y) . After that, F2 is applied to R0 (x, y) , R1 (x, y) , and neighbouring samples to derive a filtered sample as where fi, j is the clipped difference between a neighbouring sample and current sample R (x, y) and gi is the clipped difference between Ri-20 (x, y) and current sample. The filter coefficients ci, i=0, …21, are signalled.
[0049] 4. Classification
[0050] Based on directionality Di and activity aclass Ci is assigned to each 2x2 block: where MD, i represents the total number of directionalities Di.
[0051] As in VVC, values of the horizontal, vertical, and two diagonal gradients are calculated for each sample using 1-D Laplacian. The sum of the sample gradients within a 4×4 window that covers the target 2×2 block is used for classifier C0 and the sum of sample gradients within a 12×12 window is used for classifiers C1 and C2. The sums of horizontal, vertical and two diagonal gradients are denoted, respectively, as and The directionality Di is determined by comparing with a set of thresholds. The directionality D2 is derived as in VVC using thresholds 2 and 4.5. For D0 and D1, horizontal / vertical edge strength and diagonal edge strength are calculated first. Thresholds Th= [1.25, 1.5, 2, 3, 4.5, 8] are used. Edge strength is 0 if otherwise, is the maximum integer such that Edge strength is 0 if otherwise, is the maximum integer such that When i.e., horizontal / vertical edges are dominant, the Di is derived by using Table 2A; otherwise, diagonal edges are dominant, the Di is derived by using Table 2B. Table 2A. Mapping of and to Di Table 2B. Mapping of and to Di
[0052] To obtain the sum of vertical and horizontal gradients Ai is mapped to the range of 0 to n, where n is equal to 4 for and 15 for and
[0053] In an ALF_APS, up to 4 luma filter sets are signalled, each set may have up to 25 filters.
[0054] 5. Alternative 2x2 ALF classifier
[0055] Classification in ALF is extended with an additional alternative classifier. For a signalled luma filter set, a flag is signalled to indicate whether the alternative classifier is applied. Geometrical transformation is not applied to the alternative band classifier. When the band-based classifier is applied, the sum of sample values of a 2x2 luma block is calculated at first. Then the class index is calculated as below, class_index = (sum *25) >> (sample bit depth + 2) .
[0056] 6. Residual based classifier
[0057] A third classifier is based on luma residual sample values. For each 2x2 luma block, the sum of absolute values of the residual samples in a neighbouring 8x8 window is calculated, and the class index is derived as: classIdx = sum >> (sample bit depth - 4) .
[0058] The value of classIdx is in the range of 0 to 24, same as in ECM-8.0. The classifier usage is signalled for each luma filter set in APS.
[0059] 7. CCALF with long tap filter
[0060] The CCALF process uses a linear filter to filter luma sample values and generate a residual correction for the chroma samples. A 25-tap large filter is used in CCALF process, which is illustrated in Fig. 5. In Fig. 5, taps for luma samples are shown in grey dots and the location of the corresponding chroma sample is shown as a small dash-lined circle. For a given slice, the encoder can collect the statistics of the slice, analyse them and can signal up to 16 filters through APS.
[0061] 8. Adaptive filter shape switch / using samples before deblocking filter for ALF
[0062] Two candidate filter shapes: a diamond shape as shown in Fig. 6 and a new cross shape as shown in Fig. 7, can be adaptively selected by the luma filters in ALF. The number of coefficients of a luma filter is 22 for both the filter shapes. Please note that these 22 taps are constituted with 20 spatial taps (610 and 710 in Fig. 6 and Fig. 7 respectively) and 2 fixed filters based taps (620 and 720 in Fig. 6 and Fig. 7 respectively) in both shapes.
[0063] In each Adaptation Parameter Set (APS) , a shape index for the derived luma filters is signalled to the decoder. Each APS contains the luma filters that are associated with the filter shape index.
[0064] For each CTB, an APS index is signalled to indicate which luma filter shape is used to filter the current CTB. When filtering a luma sample, the coefficients and clip indices are also rearranged according to the corresponding filter shape.
[0065] The diamond shape luma ALF is replaced by the longer filter shown in Fig. 7.
[0066] The samples before deblocking filters are used as additional inputs for ALF. A final ALF sample is derived by weighting the regular ALF and the filter applied to the samples before the deblocking filter. Specifically, a filtered sample is derived as where fi, j is the clipped difference between a neighbouring sample and current sample R (x, y) , gi is the clipped difference between an intermediate sample and current sample R (x, y) and hi, j is the clipped difference between a neighbouring sample before DBF and current sample R (x, y) . The filter coefficients ci, i=0, …24 are signalled. In example, 3x3 diamond shape is applied to samples before deblocking filter. In an APS, a flag is signalled to indicate whether samples before DBF are used for ALF which is always set as true at encoder.
[0067] 9. Extended fixed-filter-output based taps for ALF
[0068] In ALF online-trained filters consist of 4 kinds of filter taps: spatial taps (810) , reconstruction-before-DBF based taps (840) , residual based taps (850) and fixed-filter-output based taps (820 and 830) as shown in Fig. 8.
[0069] 10. ALF with residual samples
[0070] The residual samples are used as additional inputs to the ALF. A filtered sample is derived as: where ri is the clipped neighbouring residual sample value and rFilteredi is the clipped residual sample filtered by the fixed-filter. For residual samples, the fixed filter reuses the offline fixed filter trained for reconstruction after SAO.
[0071] 11. Additional fixed filter for ALF
[0072] Additional fixed filter with a shape of diamond 7x7 is introduced, the filter parameters are stored at both the encoder and the decoder. There is no classification for the newly added fixed filter.
[0073] An online filter or online-trained filter of the proposed method is shown in Fig. 9, where spatial taps 910 (i.e., tap #0 ~ #19) , reconstruction-before-DBF-based taps 940 (i.e., tap #26, #27, #36) , residual-based taps 950 (i.e., #37 ~ #38) and fixed-filter-output-based taps 920 and 930 (i.e., tap #20 ~ #25, #34, #35) are kept the same as the ECM-8.0, and several extended taps 960 (i.e., tap #28 ~ #33, #39) are introduced into luma online-trained filters. The reconstruction before DBF is fed into the additional fixed filter to produce the filter outputs, then these filter outputs are used as input for newly extended taps. The online filter or online-trained filter refers to a filter specified in APS (Adaptation Parameter Set) , where the filter is trained at the encoder and signalled to decoder. The online filter or online-trained filter is in contrast to fixed filters, which are offline-trained and pre-defined in the specification.
[0074] This filter is always enabled without any filter shape switching.
[0075] 12. Improved fixed filters for ALF
[0076] Two Laplacian-based classifiers (one for each fixed filter) are applied to a 2x2 block. In each classifier, activity and directionality values are derived based on vertical, horizontal, and diagonal gradients using a window surrounding each 2x2 block. For each 2x2 block, the mean value of a surrounding window is calculated. Then, for each sample of this window, the difference between the sample value and the mean value is calculated. A scaling factor is determined based on the activity value derived from a Laplacian classifier. The square root of the sum of the squared differences is further quantized to C′ by a scaling factor. The value of C′is an integer between 0 and 7, inclusively. With i=0, 1, let Ci denote the classifier from the classifier of i-th fixed filter in ECM-9.0. Then the proposed class index Ci′is derived as C′i= C′*896+Ci.
[0077] The total number of the fixed filters is not changed.
[0078] Then a class index is determined based on the activity and directionality values. Two diamond shaped fixed filters are selected from the two filter sets by using the derived two class indices. Both fixed filters are applied to samples before DBF and ALF input, where additional diamond 9x9 filter is used for the samples before DBF. The shape of the first fixed filter applied to the ALF input samples is reduced from 13x13 to 9x9, and the shape of the second fixed filter, which is 13x13, applied to ALF input is unchanged as shown in Table 3. Table 3. Comparison of fixed filters between ECM-9.0 and JVET-AE0139
[0079] Fixed filter f1 is applied to outputs of f0 (instead of ALF input) and samples before DBF.
[0080] Finally, a signalled filter is applied to the ALF input samples, samples before the deblocking filter (DBF) , outputs of the two fixed filters, output of a Gaussian filter and the residual data.
[0081] 13. Luma Residual Taps in CCALF (JVET-AF0197)
[0082] For CCALF, five luma residual taps in a cross 3x3 shape 1020 are added to the 9x9 CCALF filter shape 1010 as shown in Fig. 10. The extended taps take the co-located and neighbouring luma residual values as input.
[0083] 14. ALF residuals scaling
[0084] A scaling factor is signalled in slice header, the scaling factors is applied to the difference between the ALF input and ALF output, and the scaled residual is added to the ALF input (it produces a scaled ALF filtering) . A similar scaling process is applied to NN filtering in NNVC.
[0085] Different luma scaling factors may be associated with different group of class indexes, and the ALF output is derived as follows: rec’ (s) = rec (s) + (corr (s) *tab [sfi [class (s) ] ] + 4 ) >> 3, where ALF residual correction ‘corr (s) ’ is scaled using the scaling factor associated to the class index of the sample, ‘tab’ is the predefined LUT for mapping scaling index ‘sfi’ to scaling factor.
[0086] 15. ALF-CCCM (JVET-AJ0173)
[0087] The proposed method of JVET-AJ0173 aims to improve the quality of the reconstructed chroma samples. A decoder-side diagram of the proposed method is illustrated in Figure 18.
[0088] In the proposed filtering scheme of JVET-AJ0173, each CTU is divided into non-overlapping blocks and for each block the cross-component filter coefficients are derived using the SAO (1110, 1132 and 1134) / CC-SAO (1120, 1122 and 1124) outputs as shown in Fig. 11. The output samples of luma ALF 1140 are used as input to the ALF-CCCM filtering 1152. The parameters for the CCCM filtering 1152 are derived using CCCM filter derivation 1150 with SAO process Y” , Cb” and Cr” . Also, the luma SAO output (i.e., Y” ) is also used as input for CC-ALLF Cb 1162 and CC-ALF Cr 1164. To obtain the final ALF-CCCM output samples, the cross-component prediction samples are blended with the SAO / CC-SAO chroma output samples. In the blending the weights are equal to 0.5 for both the SAO / CC-SAO chroma output samples and the cross-component prediction samples. The final ALF-CCCM output samples are fed to ALF chroma 1170 module to generate the adjustment signal for Cb and Cr.
[0089] The encoder decides the best block size for each CTU using a rate-distortion optimization loop. There are eight possible blocks sizes 2x2, 3x3, 4x4, 8x8, 16x16, 32x32, 64x64, 128x128. For filter derivation the blocks are extended by one sample on each side. For example, with 2x2 blocks the filter is derived using 4x4 blocks. The block extension is clipped against CTU boundaries.
[0090] For each CTU, the encoder’s RDO decides the best cross-component model from eight possible models. The models are listed in Table 4, where the cardinal directions indicate co-located luma sample position in the chroma grid (north up, south down) . The nonlinear and bias terms are the same as in the 7-tap CCCM model. The CCCM solver is used for deriving the filter coefficients. The 6-tap downsampling filter (used in CCCM and CCLM) is used for mapping the co-located luma into the chroma grid.
[0091] For each CTU, the choice of the block size and the choice of the cross-component model are signalled using CABAC coded flags.
[0092] For intra-coding, a CTU may inherit both the block size and the model type from the above or left CTU. This choice is signalled using a single CABAC coded flag for each CTU when ALF-CCCM is present in the left and / or above CTU. If two CTU candidates available, another CABAC coded flag is signalled to indicate the choice.
[0093] For inter-coding, a picture may inherit the block sizes and model types for all CTUs from a reference picture. The reference picture is derived from the L0 and L1 lists. Only reference pictures with ALF-CCCM present in at least one CTU are considered. The reference picture with the smallest POC distance to the current POC is selected. If activated, the picture level inheritance will skip the CTU-level signalling completely for the current picture. This choice is signalled using a single CABAC coded flag if a reference picture with at least one CTU utilizing ALF-CCCM is present.
[0094] The existing operations of SAO, CC-SAO, ALF and CC-ALF of ECM-14.0 are kept unchanged with exception that ALF chroma operates on ALF-CCCM output. Table 4. The eight cross-component models of ALF-CCCM
[0095] In the present invention, methods and apparatus to reduce processing latency for video coding systems utilising ALF and CCM (Cross-Component Model) are disclosed. BRIEF SUMMARY OF THE INVENTION
[0096] Method and apparatus to reduce latency for video coding systems incorporating Adaptive Loop Filter (ALF) processing and Cross-Component Model (CCM) are disclosed. According to one method, input data for a current block is received, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block. CCM (Cross-Component Model) parameters for the current block are determined. CCM process with the CCM parameters and a first process in ILF (In-Loop-Filtering) processes are applied to the input data for the current block to generate CCM output and first-process output, wherein inputs to the CCM process and the ILF process comprise no output data from the CCM process, the first process, or any other ILF process, or wherein the input to the ILF process comprises output data from the CCM process in an allowed region of the output data from the CCM process for a current sample or the current block, and the allowed region of the output data from the CCM process is restricted to a left, top or both top and left region of the current sample or the current block. A corrected output comprising the CCM output and the first-process output is derived. The corrected output or the first-process output is provided.
[0097] In the one embodiment, the first process belongs to a group comprising SAO (Sample Adaptive Offset) , CCSAO (Cross-Component SAO) , BIF (Bilateral Filter) , ALF (Adaptive Loop Filter) , CCALF (Cross-Component ALF) , or a combination thereof.
[0098] In one embodiment, the corrected output comprises a weighted sum of the CCM output and the first-process output. In one embodiment, a weighting factor for the weighted sum of the CCM output and the first-process output is signalled or pre-defined.
[0099] According to another method, input data for a current block is received, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block. Different residual scaling processes are applied to ALF process and CCALF process to generate scaled ALF output and scaled CCALF output. The scaled ALF output and the scaled CCALF output are added to a target reconstructed sample or a target processed-reconstructed sample of the current block to generate a corrected output. The corrected output is provided.
[0100] In one embodiment, residual scaling process is only applied to one of the ALF process and the CCALF process.
[0101] In one embodiment, two separate scaling factors are used for the ALF process and the CCALF process. In one embodiment, chroma classification is used to select target scaling factors used for the ALF process and the CCALF process. In one embodiment, when scaling is applied to the ALF process or the CCALF process, result of ALF or CCALF chroma classification is reused for scaling factor selection.BRIEF DESCRIPTION OF THE DRAWINGS
[0102] Fig. 1A illustrates an exemplary adaptive Inter / Intra video coding system incorporating loop processing.
[0103] Fig. 1B illustrates a corresponding decoder for the encoder in Fig. 1A.
[0104] Fig. 2 illustrates the ALF filter shapes for the chroma (left) and luma (right) components.
[0105] Figs. 3A-D illustrates the subsampled Laplacian calculations for gv (3A) , gh (3B) , gd1 (3C) and gd2 (3D) .
[0106] Fig. 4A illustrates the placement of CC-ALF with respect to other loop filters.
[0107] Fig. 4B illustrates a diamond shaped filter for the chroma samples.
[0108] Fig. 5 illustrates the 25-tap large filter used in CCALF process.
[0109] Fig. 6 illustrates the diamond shaped ALF in ECM-5.0.
[0110] Fig. 7 illustrates a longer ALF as an alternative to the diamond shaped ALF in Fig. 6.
[0111] Fig. 8 illustrates the filter shape of ALF in ECM-7.0.
[0112] Fig. 9 illustrates an example of ALF with additional fixed filter.
[0113] Fig. 10 illustrates the newly introduced filter shape for CCALF according to JVET-AF0197.
[0114] Fig. 11 illustrates decoder-side diagram of ALF-CCCM, where the CCCM filters are derived using SAO / CC-SAO outputs, and the CCCM filtering uses the ALF luma output samples as its input and resulting samples are blended with the SAO / CC-SAO outputs before ALF chroma processing.
[0115] Fig. 12 illustrates an example of ALF and CCM processes being performed in parallel to reduce processing latency according to an embodiment of the present invention.
[0116] Fig. 13 illustrates an example of BIF, SAO, CCSAO and CCM processes being performed in parallel to reduce processing latency according to an embodiment of the present invention.
[0117] Fig. 14 illustrates an example of using CCM output as input to the ALF, where the allowed region of the CCM output for a current sample / block is restricted to the left / top region of the current sample / block.
[0118] Fig. 15 illustrates a flowchart of an exemplary video coding system that performs CCM process and a process in the in-loop-filter processes in parallel or allows the process in the in-loop-filter processes to use output of the CCM process in a causal region according to an embodiment of the present invention.
[0119] Fig. 16 illustrates a flowchart of an exemplary video coding system that performs different residual clipping processes to ALF process and CCALF process according to an embodiment of the present invention.DETAILED DESCRIPTION OF THE INVENTION
[0120] It will be readily understood that the components of the present invention, as generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations. Thus, the following more detailed description of the embodiments of the systems and methods of the present invention, as represented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. References throughout this specification to “one embodiment, ” “an embodiment, ” or similar language mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present invention. Thus, appearances of the phrases “in one embodiment” or “in an embodiment” in various places throughout this specification are not necessarily all referring to the same embodiment.
[0121] Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. One skilled in the relevant art will recognize, however, that the invention can be practiced without one or more of the specific details, or with other methods, components, etc. In other instances, well-known structures, or operations are not shown or described in detail to avoid obscuring aspects of the invention. The illustrated embodiments of the invention will be best understood by reference to the drawings, wherein like parts are designated by like numerals throughout. The following description is intended only by way of example, and simply illustrates certain selected embodiments of apparatus and methods that are consistent with the invention as claimed herein.
[0122] PROPOSED METHOD
[0123] Latency Reduction for CCM and Residual Scaling in In-Loop Filtering
[0124] In ECM, some improvements are achieved by introducing additional stages to the existing in-loop filtering process. In JVET-AJ-0173, a cross-component convolutional model (CCCM) is used to refine the chroma samples before ALF as an ALF pre-processing stage, where the chroma ALF takes the CCCM-refined chroma samples as inputs. In JVET-AI-0084, a set of scaling factors is used to refine the ALF corrections as an ALF post-processing stage, where the scaled ALF outputs are the final corrections applied to the reconstructed samples. For those approaches, although there is coding benefit, some latency may be introduced to the codec, especially at the decoder. Therefore, in this invention, several in-loop filtering designs are proposed to reduce the latency.
[0125] Embodiment 1
[0126] In one embodiment, a cross-component model (CCM) is used to reduce coding artefacts of reconstructed samples in the in-loop filtering, and the CCM process can be performed in parallel with a first process in the in-loop filtering. That is, the inputs of the two processes (CCM process and the first process) do not comprise any outputs or intermediate results of the other process. The first process can be SAO, CCSAO, BIF, ALF, CCALF, or a combination thereof. A final output is derived by taking the sum or weighted sums of the outputs of CCM process and the first process.
[0127] Example 1.
[0128] CCM and ALF can be performed in parallel. The block diagram in Fig. 12 shows such example.
[0129] For a reconstructed sample r0 of one component (component A in the diagram) , ALF 1210 and CCM processes 1220 will provide two correction terms rALF and rCCM, respectively, where ALF and CCM take post-SAO samples as an input with CCM using cross-component information (component B in the diagram) . The final reconstruction of the sample is r0 + rALF +rCCM, or a linear combination of them, such as r0 + s0*rALF + s1*rCCM, where s0 and s1 are signalled or pre-defined weights.
[0130] Example 2.
[0131] CCM and BIF / SAO / CCSAO can be performed in parallel. Fig. 13 illustrates an exemplary block diagram.
[0132] For a reconstructed sample r0 of one component (component A in the diagram) , BIF 1310, SAO 1320, CCSAO 1330, and CCM 1340 processes would provide four correction terms rBIF, rSAO, rCCSAO, and rCCM, respectively, where all the processes take post-DBF (deblocking filter) samples as an input with CCSAO and CCM using cross-component information (component B in the diagram) . The final reconstruction of the sample is r0 + rBIF + rSAO + rCCSAO + rCCM, or a linear combination of them, such as r0 + s0*rBIF + s1*rSAO + s2*rCCSAO + s3*rCCM, where s0 to s3 are signalled or pre-defined weights.
[0133] Embodiment 2
[0134] In one embodiment, a Cross-Component Model (CCM) is used to reduce coding artefacts of reconstructed samples in the in-loop filtering, and the CCM output is used as the input of a first process in the in-loop filtering with spatial constraints. When a coding unit undergoes the first process, the allowed region of the CCM output is restricted to a causal region of the coding unit. A causal region of a coding unit comprises the region of the units processed before the coding unit and the region of the coding unit itself.
[0135] Example 3.
[0136] CCM output is an input of ALF. The filter shape of ALF contains taps using the CCM output, but the allowed region of the CCM output for a current sample / block is restricted to the left / top region of the current sample / block, and the region of the current sample / block.
[0137] An example of the block diagram is shown in Fig. 14. CCM 1420 output is one of the ALF 1410 inputs, and three versions of the allowed region of the CCM output for a current sample / block are illustrated in blocks 1430-1434, where the small block with black edges is the current sample / block, and the grey region is the allowed causal region.
[0138] Embodiment 3
[0139] In one embodiment, scaling is applied to a first process in the in-loop filtering with single clipping stage. Specifically, for each reconstructed sample, a correction term without clipping is provided by the first process, and the correction term is scaled by a scaling factor and added back to the reconstructed sample with clipping. The first process can be SAO, CCSAO, BIF, ALF, CCALF, or a combination thereof.
[0140] Clipping refers to the operation of making the sample value be in a valid range.
[0141] Example 4.
[0142] Scaling is applied to ALF and / or CCALF. Let r0 be the reconstructed sample before ALF and / or CCALF, rALF be the correction term provided by ALF and / or CCALF without clipping, and s be the scaling factor for the sample. The post-ALF / CCALF sample value is: Clip (r0+s*rALF, vmin, vmax) = min (max (r0+s*rALF, vmin) , vmax) , where vmin and vmax are the minimum and maximum valid values for a reconstructed sample.
[0143] It is emphasized that rALF is simply the product sum calculated in ALF / CCALF filtering process without clipping. ci is a signaled ALF coefficient and ni is an ALF filter input.
[0144] In one prior art (JVET-AI-0084) , rALF is clipped for both luma and chroma. For luma, the clipping is applied as follows: and for chroma, the clipping is applied as follows:
[0145] The clipping range used in this prior art is determined by the input bit-depth: vmax=-vmin=2input_bitdepth.
[0146] The clipped correction r′ALF is used to generate the final correction with the scaling factor s, where another clipping is performed: Clip (r0+s*r′ALF, vmin, vmax) = min (max (r0+s*r′ALF, vmin) , vmax)
[0147] Compared to the prior art, the proposed design requires a shorter critical path to process a reconstructed sample due to the reduction of clipping operations.
[0148] Embodiment 4
[0149] In one embodiment, scaling is applied to either ALF or CCALF, but not both. Specifically, for each reconstructed sample, two correction terms, rALF and rCCALF, are provided by ALF and CCALF, respectively. However, only one of the rALF and rCCALF is scaled by a scaling factor and added back to the reconstructed sample.
[0150] Embodiment 5
[0151] In one embodiment, scaling is applied to ALF and CCALF with separate scaling factors. Specifically, for each reconstructed sample, two correction terms, rALF and rCCALF, are provided by ALF and CCALF, respectively. rALF is scaled by a first scaling factor and rCCALF is scaled by a second scaling factor. The scaled correction terms are added back to the reconstructed sample.
[0152] Embodiment 6
[0153] This embodiment is based on embodiment 4 or 5. However, chroma classification is used to select the scaling factor.
[0154] Embodiment 7
[0155] In embodiment 6, when scaling is applied to ALF, the ALF chroma classification result is reused for the scaling factor selection; when scaling is applied to CCALF, the CCALF chroma classification result is reused for the scaling factor selection.
[0156] The foregoing proposed methods can be implemented in encoders and / or decoders. For example, the proposed method can be implemented in an in-loop filtering module of an encoder, and / or an in-loop filtering module of a decoder. For example, any of the proposed methods can be implemented in the in-loop filter module (e.g. ILPF 130 in Fig. 1A and Fig. 1B) of an encoder or a decoder. Alternatively, any of the proposed methods can be implemented as circuits coupled to the inter coding module of an encoder and / or motion compensation module, a merge candidate derivation module of the decoder. The methods may also be implemented using executable software or firmware codes stored on a media, such as hard disk or flash memory, for a CPU (Central Processing Unit) or programmable devices (e.g. DSP (Digital Signal Processor) or FPGA (Field Programmable Gate Array) ) .
[0157] Fig. 15 illustrates a flowchart of an exemplary video coding system that performs CCM process and a process in the in-loop-filter processes in parallel or allows the process in the in-loop-filter processes to use output of the CCM process in a causal region according to an embodiment of the present invention. The steps shown in the flowchart may be implemented as program codes executable on one or more processors (e.g., one or more CPUs) at the encoder side. The steps shown in the flowchart may also be implemented based hardware such as one or more electronic devices or processors arranged to perform the steps in the flowchart. According to the method, input data for a current block is received in step 1510, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block. CCM (Cross-Component Model) parameters for the current block are determined in step 1520. CCM process with the CCM parameters and a first process in ILF (In-Loop-Filtering) processes are applied to the input data for the current block to generate CCM output and first-process output in step 1530, wherein inputs to the CCM process and the ILF process comprise no output data from the CCM process, the first process, or any other ILF process, or wherein the input to the ILF process comprises output data from the CCM process in an allowed region of the output data from the CCM process for a current sample or the current block, and the allowed region of the output data from the CCM process is restricted to a left, top or both top and left region of the current sample or the current block. A corrected output comprising the CCM output and the first-process output is derived in step 1540. The corrected output or the first-process output is provided in step 1550.
[0158] Fig. 16 illustrates a flowchart of an exemplary video coding system that performs different residual clipping processes to ALF process and CCALF process according to an embodiment of the present invention. According to another method, input data for a current block is received in step 1610, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block. Different residual scaling processes are applied to ALF process and CCALF process to generate scaled ALF output and scaled CCALF output in step 1620. The scaled ALF output and the scaled CCALF output are added to a target reconstructed sample or a target processed-reconstructed sample of the current block to generate a corrected output in step 1630. The corrected output is provided in step 1640.
[0159] The flowchart shown is intended to illustrate an example of video coding according to the present invention. A person skilled in the art may modify each step, re-arranges the steps, split a step, or combine steps to practice the present invention without departing from the spirit of the present invention. In the disclosure, specific syntax and semantics have been used to illustrate examples to implement embodiments of the present invention. A skilled person may practice the present invention by substituting the syntax and semantics with equivalent syntax and semantics without departing from the spirit of the present invention.
[0160] The above description is presented to enable a person of ordinary skill in the art to practice the present invention as provided in the context of a particular application and its requirement. Various modifications to the described embodiments will be apparent to those with skill in the art, and the general principles defined herein may be applied to other embodiments. Therefore, the present invention is not intended to be limited to the particular embodiments shown and described, but is to be accorded the widest scope consistent with the principles and novel features herein disclosed. In the above detailed description, various specific details are illustrated in order to provide a thorough understanding of the present invention. Nevertheless, it will be understood by those skilled in the art that the present invention may be practiced.
[0161] Embodiment of the present invention as described above may be implemented in various hardware, software codes, or a combination of both. For example, an embodiment of the present invention can be one or more circuit circuits integrated into a video compression chip or program code integrated into video compression software to perform the processing described herein. An embodiment of the present invention may also be program code to be executed on a Digital Signal Processor (DSP) to perform the processing described herein. The invention may also involve a number of functions to be performed by a computer processor, a digital signal processor, a microprocessor, or field programmable gate array (FPGA) . These processors can be configured to perform particular tasks according to the invention, by executing machine-readable software code or firmware code that defines the particular methods embodied by the invention. The software code or firmware code may be developed in different programming languages and different formats or styles. The software code may also be compiled for different target platforms. However, different code formats, styles and languages of software codes and other means of configuring code to perform the tasks in accordance with the invention will not depart from the spirit and scope of the invention.
[0162] The invention may be embodied in other specific forms without departing from its spirit or essential characteristics. The described examples are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is therefore, indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.
Claims
1.A method for in-loop filtering of reconstructed video data, the method comprising:receiving input data for a current block, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block;determining CCM (Cross-Component Model) parameters for the current block;applying CCM process with the CCM parameters and a first process in ILF (In-Loop-Filtering) processes to the input data for the current block to generate CCM output and first-process output, wherein inputs to the CCM process and the ILF process comprise no output data from the CCM process, the first process, or any other ILF process, or wherein the input to the ILF process comprises output data from the CCM process in an allowed region of the output data from the CCM process for a current sample or the current block, and the allowed region of the output data from the CCM process is restricted to a left, top or both top and left region of the current sample or the current block;deriving a corrected output comprising the CCM output and the first-process output; andproviding the corrected output or the first-process output.2.The method of Claim 1, wherein the first process belongs to a group comprising SAO (Sample Adaptive Offset) , CCSAO (Cross-Component SAO) , BIF (Bilateral Filter) , ALF (Adaptive Loop Filter) , CCALF (Cross-Component ALF) , or a combination thereof.3.The method of Claim 1, wherein the corrected output comprises a weighted sum of the CCM output and the first-process output.4.The method of Claim 3, wherein a weighting factor for the weighted sum of the CCM output and the first-process output is signalled or pre-defined.5.An apparatus for video coding, the apparatus comprising one or more electronics or processors arranged to:receive input data for a current block, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block;determine CCM (Cross-Component Model) parameters for the current block;apply CCM process with the CCM parameters and a first process in ILF (In-Loop-Filtering) processes to the input data for the current block to generate CCM output and first-process output, wherein inputs to the CCM process and the ILF process comprise no output data from the CCM process, the first process, or any other ILF process, or wherein the input to the ILF process comprises output data from the CCM process in an allowed region of the output data from the CCM process for a current sample or the current block, and the allowed region of the output data from the CCM process is restricted to a left, top or both top and left region of the current sample or the current block;derive a corrected output comprising the CCM output and the first-process output; andprovide the corrected output or the first-process output.6.A method for in-loop filtering of reconstructed video data, the method comprising:receiving input data for a current block, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block;applying different residual scaling processes to ALF process and CCALF process to generate scaled ALF output and scaled CCALF output;adding the scaled ALF output and the scaled CCALF output to a target reconstructed sample or a target processed-reconstructed sample of the current block to generate a corrected output; andproviding the corrected output.7.The method of Claim 6, wherein residual scaling process is only applied to one of the ALF process and the CCALF process.8.The method of Claim 6, wherein two separate scaling factors are used for the ALF process and the CCALF process.9.The method of Claim 8, wherein chroma classification is used to select target scaling factors used for the ALF process and the CCALF process.10.The method of Claim 8, wherein when scaling is applied to the ALF process or the CCALF process, result of ALF or CCALF chroma classification is reused for scaling factor selection.11.An apparatus for video coding, the apparatus comprising one or more electronics or processors arranged to:receive input data for a current block, wherein the input data comprises reconstructed samples or processed-reconstructed samples of the current block;apply different residual scaling to ALF process and CCALF process to generate scaled ALF output and scaled CCALF output;add the scaled ALF output and the scaled CCALF output to a target reconstructed sample or a target processed-reconstructed sample of the current block to generate a corrected output; andprovide the corrected output.