METHOD AND APPARATUS FOR CHROMA PREDICTION IN VIDEO CODING SYSTEMS

Abstract

A method and apparatus for deriving the cross-component prediction using reference data, for example, compensated prediction or template reconstruction, of corresponding luma and chroma components. According to this method, whether a target mode is applied to the current block is determined. If the target mode is applied to the current block, the following steps are performed. A target candidate cross-component predictor for the second-colour block is derived where a target candidate cross-component model associated with the target candidate cross-component predictor is derived using reference data of corresponding first-colour component and corresponding second-colour component for the current block. The reference data is associated with a reference region comprising template of the current block or a pre-defined region indicated using a vector. A final prediction is derived using the target candidate cross-component predictor. The second-colour block is encoded or decoded using the final prediction.

Description

1 / 94 METHOD AND APPARATUS FOR CHROMA PREDICTION IN VIDEO CODING SYSTEMS CROSS-REFERENCE TO RELATED REQUESTS

[0001] The present invention is a non-provisional application and claims priority over U.S. Provisional Patent Application No. 63 / 490,806, filed March 17, 2023, over U.S. Provisional Patent Application No. 63 / 497,758, filed April 24, 2023, and over U.S. Provisional Patent Application No. 63 / 498,853, filed April 28, 2023. The U.S. Provisional Patent Applications are incorporated herein by reference in their entirety. FIELD OF THE INVENTION

[0002] The present invention relates to a video encoding system. In particular, the present invention relates to schemes for improving chroma prediction through cross-component prediction derivation. For example, compensated prediction generated using motion information and / or block vectors, or model reconstruction, of corresponding luma and chroma components is used to derive cross-component prediction. BACKGROUND AND STATE OF THE RELATED ART

[0003] Versatile Video Coding (VVC) is the latest international video coding standard developed by the Joint Video Expert Team (JVET) of the ITU-T Video Coding Expert Group (VCEG) and the ISO / IEC Moving Picture Expert Group (MPEG). The standard was published as an ISO standard: ISO / IEC 23090-3:2021, Information technology – Encoded representation of immersive media – Part 3: Versatile video coding, published in February 2021. VVC was developed based on its predecessor, HEVC (High Efficiency Video Coding), adding more coding tools to improve coding efficiency and also to handle various types Petition 870250094644, dated 10 / 16 / 2025, page 9 / 111 2 / 94 of video sources, including three-dimensional (3D) video signals.

[0004] Figure 1A illustrates an exemplary Inter or Intra adaptive video coding system incorporating loop processing. For Intra prediction, prediction data is derived based on previously encoded video data in the current image. For Inter prediction 112, motion estimation (ME) is performed on the encoder side and motion compensation (MC) is performed based on the ME result to provide prediction data derived from other images and motion data. Switch 114 selects intra prediction 110 or inter prediction 112 and the selected prediction data is provided to the adder 116 to form prediction errors, also called residuals. The prediction error is then processed by transformation (T) 118, followed by quantization (Q) 120.The transformed and quantized residues are then encoded by Entropy Encoder 122 to be included in a video bitstream corresponding to the compressed video data. The bitstream associated with the transformation coefficients is then compressed with secondary information, such as motion and encoding modes associated with Intra prediction and Inter prediction, and other information, such as parameters associated with loop filters applied to the underlying image area. The secondary information associated with intra prediction 110, inter prediction 112, and loop filter 130 is provided to Entropy Encoder 122, as shown in Figure 1A. When an inter prediction mode is used, one or more reference images also need to be reconstructed at the encoder end. Consequently, the transformed and quantized residues are processed by Inverse Quantization (IQ) 124 and Inverse Transformation (IT) 126 to recover the residues.The residuals are then added back to the prediction data. Petition 870250094644, dated 10 / 16 / 2025, page 10 / 111 3 / 94 136 in Reconstruction (REC) 128 to reconstruct the video data. The reconstructed video data can be stored in the Reference Image Buffer 134 and used for the prediction of other frames.

[0005] As shown in Figure 1A, the received video data undergoes a series of processing steps in the encoding system. The video data reconstructed from REC 128 may be subject to various deteriorations due to a series of processing steps. Thus, loop filtering 130 is frequently applied to the reconstructed video data before this data is stored in the Reference Image Buffer 134 in order to improve video quality. For example, unlocking filtering (DF), adaptive sample shifting (SAO), and adaptive loop filtering (ALF) may be used. Loop filter information may need to be incorporated into the bitstream so that a decoder can correctly recover the necessary information. Therefore, loop filter information is also provided to the entropy encoder 122 for incorporation into the bitstream.In Figure 1A, loop filter 130 is applied to the reconstructed video before the reconstructed samples are stored in the reference image buffer 134. The system in Figure 1A aims to illustrate an exemplary structure of a typical video encoder. It may correspond to the High Efficiency Video Coding (HEVC), VP8, VP9, ​​H.264, or VVC system.

[0006] The decoder, as shown in Figure 1B, can use similar functional blocks or parts of the same functional blocks as the encoder, except for Transformer 118 and Quantizer 120, since the decoder only needs Inverse Quantizer 124 and Inverse Transformer 126. Instead of Entropy Encoder 122, the decoder uses Entropy Decoder 140 to decode the video bitstream into quantized transformation coefficients and Petition 870250094644, dated 10 / 16 / 2025, page 11 / 111 4 / 94 Required encoding information (e.g., ILPF information, intra prediction information, and inter prediction information). The Intra prediction 150 on the decoder side does not need to perform mode search. Instead, the decoder only needs to generate the Intra prediction according to the Intra prediction information received from the Entropy Decoder 140. Furthermore, for the Inter prediction, the decoder only needs to perform motion compensation (MC 152) according to the Inter prediction information received from the Entropy Decoder 140, without the need for motion estimation.

[0007] According to VVC, an input image is divided into non-overlapping square block regions, called CTUs (Coding Tree Units), similar to HEVC. Each CTU can be divided into one or more smaller coding units (CUs). The resulting CU partitions can have square or rectangular shapes. Furthermore, VVC divides a CTU into prediction units (PUs) as a unit to apply the prediction process, such as inter-prediction, intra-prediction, etc. Partitioning of CTUs using a tree structure

[0008] In VVC, a quadtree with nested multitype trees using binary and tertiary split segmentation structure replaces the concepts of various partition unit types, i.e., it removes the separation of the concepts of CU, PU, ​​and TU, except when necessary for CUs that have a size too large for the maximum transformation length, and offers more flexibility for CU partition formats. In the coding tree structure, a CU can have a square or rectangular shape. A coding tree unit (CTU) is first partitioned by a quaternary tree structure (also known as a quadtree). Then, the leaf nodes of the quaternary tree can be further partitioned by a multitype tree structure. In most cases, the CU, PU, ​​and TU have Petition 870250094644, dated 10 / 16 / 2025, page 12 / 111 5 / 94 the same block size in the quad tree with nested multi-type tree encoding block structure. The exception occurs when the maximum supported transformation length is less than the width or height of the CU color component.

[0009] In VVC, the coding tree scheme supports the ability for luminance and chrominance to have separate block tree structures. For P and B slices, the luma and chroma CTBs in a CTU must share the same coding tree structure. However, for I slices, luma and chroma can have separate block tree structures. When the separate block tree mode is applied, the luma CTB is partitioned into CUs by one coding tree structure, and the chroma CTBs are partitioned into chroma CUs by another coding tree structure. This means that a CU in an I slice can consist of one coding block of the luma component or coding blocks of two chroma components, and a CU in a P or B slice always consists of coding blocks of all three color components, unless the video is monochrome. Intra-mode coding with 67 Intra prediction modes

[00010] To capture the arbitrary edge directions presented in natural videos, the number of intra-directional modes in VVC was expanded from 33, used in HEVC, to 65.

[00011] In VVC, several conventional intra-angle prediction modes are adaptively replaced by wide-angle intra-angle prediction modes for non-square blocks.

[00012] To keep the complexity of generating the list of most probable modes (MPMs) low, an intra-coding method with 6 MPMs is used, considering two available neighboring intra-modes. The following three aspects are considered to construct the MPM list: - Standard intra modes - Intra-neighbor modes Petition 870250094644, dated 10 / 16 / 2025, page 13 / 111 6 / 94 - Intra-derived modes.

[00013] Secondary MPM lists are introduced as described in JVET-D0114 (Seregin, et al., “Intra-mode dependent block shape coding, ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29 / WG 11 Joint Video Exploration Team (JVET), 4th Meeting: Chengdu, CN, October 15–21, 2016, JVET-D0114 document).The remaining inputs are composed of the intra modes of the neighboring blocks to the left (L), above (A), below left (BL), above right (AR) and above left (AL), the directional modes with added deviation of the first two available directional modes of the neighboring blocks, and the standard modes. DECODER SIDE INTRA MODE DERIVATION (DIMD)

[00014] When DIMD is applied, several intra modes (e.g., mode 1 and mode 2 suggested by the DIMD derivation process) are derived from the reconstructed neighboring samples, and these predictors are combined with the planar mode predictor with weights derived from the gradients. The DIMD mode is used as an alternative prediction mode and is always checked in the high-complexity RDO mode.

[00015] To implicitly derive the intra-block prediction modes, a texture gradient analysis is performed on both the encoder and decoder sides. This process begins with an empty gradient histogram (HoG) with 65 entries, corresponding to the 65 angular modes. The amplitudes of these entries are determined during the texture gradient analysis. Petition 870250094644, dated 10 / 16 / 2025, page 14 / 111 7 / 94

[00016] In the first step, DIMD selects a model with T=3 columns and rows, respectively, from the left and top sides of the current block. This area is used as a reference for deriving the gradient-based intra-prediction modes.

[00017] In the second step, horizontal and vertical Sobel filters are applied to all positions in the 3^3 window, centered on the pixels of the model's midline. At each window position, the Sobel filters calculate the intensity of the pure horizontal and vertical directions as Gx and Gy, respectively. Then, the window texture angle is calculated as: angle = arctan (Gx / Gy) (1) which can be converted into one of 65 intra-angular prediction modes. Since the index of the current window's intra-window prediction mode is derived as idx, the amplitude of its input in HoG[idx] is updated by adding: amplitude = |Gx| + |Gy| (2) .

[00018] Figures 2A-C show an example of HoG, calculated after applying the above operations to all pixel positions in the model. Figure 2A illustrates an example of the selected model 220 for a current block 210. The model 220 comprises T rows above the current block and T columns to the left of the current block. For intra-current block prediction, the area 230 above and to the left of the current block corresponds to a reconstructed area and the area 240 below and to the right of the block corresponds to an unavailable area. Figure 2B illustrates an example for T=3 and the HoGs are calculated for pixels 260 in the middle row and pixels 262 in the middle column. For example, for pixel 252, a 3x3 window 250 is used. Figure 2C illustrates an example of the amplitudes (ampl) calculated based on equation (2) for the intra-angular prediction modes, as determined by equation (1).

[00019] Once the HoG is calculated, if two prediction modes Petition 870250094644, dated 10 / 16 / 2025, p. 15 / 111 8 / 94 intra are derived, the indices with the two highest bars of the histogram are selected as the two implicitly derived intra prediction modes for the block and are subsequently combined with the Planar mode as the DIMD mode prediction. Prediction fusion is applied as a weighted average of the three predictors above. For this purpose, the weighting of the planar is fixed at 21 / 64 (~1 / 3). The remaining weighting of 43 / 64 (~2 / 3) is then divided between the two HoG IPMs, proportionally to the amplitude of their HoG bars. Figure 3 illustrates an example of the blending process. As shown in Figure 3, two intra modes (M1 312 and M2 314) are selected according to the indices with the two highest bars of the histogram bars 310. The three predictors (340, 342 and 344) are used to form the blended prediction.The three predictors correspond to the application of the intra modes M1, M2, and planar (320, 322, and 324, respectively) to the reference pixels 330 to form the respective predictors. The three predictors are weighted by their respective weighting factors (ω1, ω2, and ω3) 350. The weighted predictors are summed using the adder 352 to generate the combined predictor 360. Note that if only one mode (i.e., single mode) exists in the histogram, then there will be no combination process and no second DIMD mode.

[00020] In addition, the two intra-derived modes are implicitly included in the MPM list so that the DIMD process is executed before the MPM list is constructed. The primary intra-derived mode of a DIMD block is stored with a block and is used for the construction of the MPM list of neighboring blocks. Intra-Mode Derivation Based on Model (TIMD)

[00021] The intra-mode model-based derivation (TIMD) implicitly derives the intra-mode prediction of a CU using a neighboring model in both the encoder and decoder, rather than signaling the intra-mode prediction to Petition 870250094644, dated 10 / 16 / 2025, page 16 / 111 9 / 94 the decoder. As shown in Figure 4, the model prediction samples (412 and 414) for the current block 410 are generated using the model reference samples (420 and 422) for each candidate mode. A cost is calculated as the SATD (Sum of Absolute Transformed Differences) between the prediction samples and the model reconstruction samples. The intra-prediction mode with the minimum cost is selected as the DIMD mode and used for intra-prediction of the CU. The candidate modes can be 67 intra-prediction modes, as in VVC, or extended to 131 intra-prediction modes. In general, MPMs can provide a clue to indicate the directional information of a CU. Thus, to reduce the intra-mode search space and utilize the characteristics of a CU, the intra-prediction mode can be implicitly derived from the MPM list.

[00022] For each intra prediction mode in the MPMs, the SATD (TIMD cost) between prediction samples and model reconstruction is calculated. The first two intra prediction modes with the minimum SATD are selected as the TIMD modes suggested by the TIMD derivation process. These two TIMD modes are merged with the weights after applying the PDPC process, and this weighted intra prediction is used to encode the current CU. The combination of position-dependent intra prediction (PDPC) is included in the derivation of the TIMD modes.

[00023] The costs of the two selected modes are compared against a threshold. In the test, the cost factor 2 is applied as follows: costMode2 < 2*costMode1.

[00024] If this condition is true, the merger is applied; otherwise, only mode 1 is used (i.e., single-mode case). The mode weights are calculated from their SATD costs as follows: weighting1 = costMode2 / (costMode1+ costMode2) weighting2 = 1 - weighting1. Petition 870250094644, dated 10 / 16 / 2025, page 17 / 111 10 / 94 Model Correspondence Prediction (TMP)

[00025] Model matching prediction (TMP) is a special intra-prediction mode that copies the best prediction block from the reconstructed part of the current frame, whose L-shaped model matches the current model. CCLM (Crossed Component Linear Model)

[00026] The main idea behind CCLM mode (sometimes abbreviated as LM mode) is as follows: the chroma components of a block can be predicted from the co-located reconstructed luma samples by linear models whose parameters are derived from already reconstructed luma and chroma samples that are adjacent to the block.

[00027] In VVC, the CCLM mode uses interchannel dependencies, predicting chroma samples from reconstructed luma samples. This prediction is performed using a linear model in the form: P(i,j) = a rec'L(i,j) + b. (3).

[00028] Here, P(i,j) represents the chroma samples predicted in a CU and rec'L(i,j )) represents the luma samples reconstructed from the same CU that are subsampled for the non-4:4:4 color format case. The parameters of the aeb model are derived based on the reconstructed neighboring luma and chroma samples, both on the encoder and decoder sides, without explicit signaling.

[00029] Three CCLM modes, namely CCLM_LT, CCLM_L, and CCLM_T, are specified in the VVC. These three modes differ in the locations of the reference samples used for deriving the model parameters. Only upper bound samples are involved in the CCLM_T mode, and only left bound samples are involved in the CCLM_L mode. In the CCLM_LT mode, both upper and left bound samples are used. MMLM OVERVIEW Petition 870250094644, dated 10 / 16 / 2025, page 18 / 111 11 / 94

[00030] As indicated by the name, the original CCLM mode employs a linear model to predict chroma samples from luma samples for the entire CU, while in MMLM (Multiple Model CCLM), two models can exist. In MMLM, neighboring luma samples and neighboring chroma samples of the current block are classified into two groups, each group is used as a training set to derive a linear model (i.e., specific α and β are derived for a specific group). Furthermore, samples from the current luma block are also classified based on the same rule for classifying neighboring luma samples. - The limit is calculated as the average value of the reconstructed neighboring luma samples. A neighboring sample with Rec'L[x,y] <= Limit is classified in group 1; while a neighboring sample with RecL[x,y] > Limit is classified in group 2. Correspondingly, a prediction for chrominance is obtained using linear models: íPredc[%,y] = arx Rec'L[x,y] + βγif Rec'L[x,y] < Limit kPredc[x,y] = a2x Rec'L[x,y] + β2if Rec'L[x,y] > Limit. CROSS-COMPONENT CONVOLUTIONAL MODEL (CCCM)

[00031] In CCCM, a convolutional model is applied to improve chroma prediction performance. The convolutional model has a 7-touch filter composed of a spatial component in the form of a signal of more than 5 touches, a nonlinear term, and a polarization term.

[00032] The filter output is calculated as a convolution between the filter coefficients and the input values ​​and clipped to the range of valid chroma samples: The filter coefficients are calculated by minimizing the MSE between the predicted and reconstructed chroma samples in the reference area.

[00033] MSE minimization is performed by calculating the autocorrelation matrix for the luma input and a cross-correlation vector between the luma input and the chroma output. Petition 870250094644, dated 10 / 16 / 2025, page 19 / 111 12 / 94 The autocorrelation matrix is ​​decomposed into LDL, and the final filter coefficients are calculated using back substitution. The process roughly follows the calculation of the ALF filter coefficients in ECM, but LDL decomposition was chosen instead of Cholesky decomposition to avoid the use of square root operations. GRADIENT LINEAR MODEL (GLM)

[00034] Compared to CCLM, instead of undersampled luminance values, GLM uses luminance sample gradients to derive the linear model. Specifically, when GLM is applied, the input to the CCLM process, i.e., the undersampled luminance samples L, are replaced by luminance sample gradients G. The other parts of CCLM (e.g., parameter derivation, prediction sample linear transformation) remain unchanged. C = α-G + β .

[00035] For signaling, when CCLM mode is enabled for the current CU, two indicators are signaled separately for the Cb and Cr components to indicate whether GLM is enabled for each component. If GLM is enabled for a component, an additional syntax element is signaled to select one of the 16 gradient filters (510-540 in Figure 5) for gradient calculation. GLM can be combined with existing CCLM by signaling an extra indicator in the bitstream. When this combination is applied, the filter coefficients used to derive the input luma samples of the linear model are calculated as the combination of the selected gradient filter from GLM and the subsampling filter from CCLM. DM Chroma Mode

[00036] For Chroma DM mode, the intra-luma block prediction mode corresponding (colocalized) that covers the center position of the current chroma block is inherited directly. COPY INTRA BLOCK Petition 870250094644, dated 10 / 16 / 2025, p. 20 / 111 13 / 94

[00037] Intrablock Copy (IBC) is a tool adopted in HEVC extensions in SCC (Screen Content Coding). It is known to significantly improve the encoding efficiency of screen content materials. As the IBC mode is implemented as a block-level encoding mode, block matching (BM) is performed in the encoder to find the ideal block vector (or motion vector) for each CU. Here, a block vector is used to indicate the offset of the current block to a reference block, which has already been reconstructed within the current image. The luma block vector of an IBC-encoded CU has integer precision. The chroma block vector is also rounded to integer precision. When combined with AMVR (Adaptive Motion Vector Resolution), the IBC mode can switch between motion vector accuracies of 1 pel and 4 pel. An IBC-encoded CU is treated as the third prediction mode, in addition to the intra and inter prediction modes.The IBC mode is applicable to CUs with a width and height less than or equal to 64 luma samples. Direct Block Vector (DBV) Mode for Chroma Prediction

[00038] The direct block vector is used for chroma blocks. An indicator is triggered to show whether a chroma block is encoded using IBC mode. If one of the luma blocks in the predefined locations is encoded with IBC or intraTMP mode, its block vector will be scaled and used as the block vector for the chroma block. Model matching is used to perform block vector scaling. INTER PREDICTION OVERVIEW

[00039] According to section 3.4 of JVET-T2002. (Jianle Chen, et al., “Description of the algorithm for versatile video coding and test model 11 (VTM 11), Joint Video Expert Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29, 20th Meeting, by teleconference, October 7-16, 2020, Document: JVET-T2002)), for each inter-planned CU, the Petition 870250094644, dated 10 / 16 / 2025, page 21 / 111 14 / 94 motion parameters consist of motion vectors, reference image indices and reference image list usage index, and additional information necessary for the new VVC coding feature to be used for interpredicted sample generation. The motion parameter can be signaled explicitly or implicitly. When a CU is coded with the jump mode, the CU is associated with a PU and has no significant residual coefficients, nor a delta-coded motion vector or reference image index. A merge mode is specified whereby the motion parameters for the current CU, which are obtained from neighboring CUs, including spatial and temporal candidates, and additional schedules introduced in the VVC, are combined. The merge mode can be applied to any interpredicted CU, not just the jump mode.The alternative to the merging mode is the explicit transmission of motion parameters, where the motion vector, the corresponding reference image index for each list of reference images, and the indicator of use of the list of reference images and other necessary information are explicitly signaled for each CU.

[00040] In addition to the inter-coding capabilities in HEVC, VVC includes a number of new and refined inter-prediction coding tools, listed below: Extended merge prediction - Merge mode with MVD (MMVD) - Symmetrical MVD signaling (SMVD) Prediction with affine motion compensation - Subblock-based temporal motion vector prediction (SbTMVP) Adaptive motion vector resolution (AMVR) - Motion field storage: 1 / 16 MV storage of the luminance sample and 8x8 motion field compression. Petition 870250094644, dated 10 / 16 / 2025, page 22 / 111 15 / 94 - Bi-prediction with weighting at the CU level (BCW) - Bidirectional optical flow (BDOF) - Decoder-side motion vector refinement (DMVR) - Geometric Partitioning Mode (GPM) - Combined inter- and intra-interference prediction (CIIP).

[00041] The following description provides details of these inter-specified prediction methods in the VVC. PREDICTION OF EXTENDED FUSION

[00042] In VVC, the list of merger candidates is constructed by including the following five types of candidates, in this order: 1. Space MVP of neighboring space CUs 2. Temporal MVP of colocalized CUs 3. MVP based on the history of a FIFO table. 4. Average MVP by peers 5. MVs zero.

[00043] The size of the merge list is signaled in the sequence parameter set (SPS) header, and the maximum allowed merge list size is 6. For each CU encoded in merge mode, an index of the best merge candidate is encoded using truncated unary binarization (TU). The first file in the merge index is context encoded, and bypass encoding is used for the remaining files.

[00044] The process for deriving each category of merge candidates is provided in this section. As with HEVC, VVC also supports parallel derivation of merge candidate lists (or so-called merge candidate lists) for all CUs within a given area size. DERIVATION OF SPACE CANDIDATES

[00045] The derivation of space fusion candidates in VVC is the same as in HEVC, except that the positions of the first two fusion candidates are swapped. A maximum of four candidates. Petition 870250094644, dated 10 / 16 / 2025, p. 23 / 111 16 / 94 The merge (B0, A0, B1, and A1) for the current CU 610 is selected from the candidates located in the positions illustrated in Figure 6A. The derivation order is B0, A0, B1, A1, and B2. Position B2 is considered only when one or more neighboring CUs of positions B0, A0, B1, and A1 are unavailable (e.g., belonging to another slice or block) or are internally encoded. After adding the candidate at position A1, the addition of the remaining candidates is subject to a redundancy check, which ensures that candidates with the same move information are excluded from the list, thus improving encoding efficiency. To reduce computational complexity, not all possible candidate pairs are considered in the aforementioned redundancy check.Instead, only pairs linked by an arrow in Figure 6B are considered, and a candidate is only added to the list if the corresponding candidate used for redundancy checking does not have the same movement information.

[00046] During the development of the VVC standard, a coding tool called Non-Adjacent Motion Vector Prediction (NAMVP) was proposed in JVET-L0399 (Yu Han, et al., “CE4.4.6: Improvement in Merge or Skip Mode, Joint Video Exploration Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29 / WG 11, 12th Meeting: Macau, CN, October 3-12, 2018, Document: JVET-L0399). According to the NAMVP technique, non-adjacent spatial merge candidates are inserted after the TMVP (i.e., the temporal MVP) in the regular list of merge candidates. The pattern of non-adjacent spatial merge candidates is shown in Figure 7. The distances between the non-adjacent spatial candidates and the current coding block are based on the width. and the height of the current coding block. In Figure 7, each small square corresponds to a NAMVP candidate, and the candidates are ordered (as shown by the number inside the square) according to distance. A Petition 870250094644, dated 10 / 16 / 2025, page 24 / 111 The 17 / 94 row buffer restriction is not applied. In other words, NAMVP candidates far from a current block may need to be stored, which could require a large buffer. DERIVATION OF TEMPORARY CANDIDATES

[00047] At this stage, only one candidate is added to the list. Specifically, in the derivation of this temporal fusion candidate for a current CU 810, a scaled motion vector is derived based on the colocalized CU 820 belonging to the colocalized reference image, as shown in Figure 8. The list of reference images and the reference index to be used for the derivation of the colocalized CU are explicitly signaled in the slice header. The scaled motion vector 830 for the temporal fusion candidate is obtained as illustrated by the dashed line in Figure 8, which is scaled from the motion vector 840 of the colocalized CU using the POC (Image Order Count) distances, tb and td, where tb is defined as the POC difference between the reference image of the current image and the current image, and td is defined as the POC difference between the reference image of the colocalized image and the colocalized image.The reference image index of the candidate for temporal fusion is set to zero.

[00048] The position for the temporal candidate is selected from candidates C0 and C1, as illustrated in Figure 9. If the CU at position C0 is unavailable, internally encoded, or outside the current line of CTUs, position C1 will be used. Otherwise, position C0 will be used in the derivation of the temporal fusion candidate. DERIVATION OF MERGER CANDIDATE BASED ON HISTORY

[00049] History-based MVP (HMVP) merge candidates are added to the merge list after the spatial MVP and TMVP. In this method, the movement information of a previously encoded block is stored in a table and used as the MVP for the current CU. The table with multiple Petition 870250094644, dated 10 / 16 / 2025, page 25 / 111 18 / 94 HMVP candidates are maintained during the encoding or decoding process. The table is reset (emptied) when a new CTU row is found. Whenever there is an unencoded CU between sub-blocks, the associated movement information is added to the last entry in the table as a new HMVP candidate.

[00050] The HMVP table size is set to 6, indicating that up to 5 history-based MVP (HMVP) candidates can be added to the table. When inserting a new movement candidate into the table, a strict first-in, first-out (FIFO) rule is used, in which redundancy checking is applied first to find out if an identical HMVP exists in the table. If found, the identical HMVP is removed from the table and all subsequent HMVP candidates are moved forward, and the identical HMVP is inserted into the last entry of the table.

[00051] HMVP candidates can be used in the process of building the merge candidate list. The last HMVP candidates in the table are checked in order and inserted into the candidate list after the TMVP candidate. Redundancy checking is applied to HMVP candidates for spatial or temporal merge candidates.

[00052] To reduce the number of redundancy check operations, the following simplifications are introduced: 1. The last two entries in the table are checked for redundancy for the spatial candidates A1 and B1, respectively. 2. When the total number of available merger candidates reaches the maximum allowed number of merger candidates minus 1, the process of building the merger candidate list from the HMVP is terminated. DERIVATION OF CANDIDATES TO PAIR-AVERAGE MERGER

[00053] Average peer candidates are generated by Petition 870250094644, dated 10 / 16 / 2025, p. 26 / 111 19 / 94 average of predefined candidate pairs in the existing merge candidate list, using the first two merge candidates. The first merge candidate is defined as p0Cand and the second merge candidate can be defined as p1Cand, respectively. Average motion vectors are calculated according to the availability of the motion vector of p0Cand and p1Cand separately for each reference list. If both motion vectors are available in a list, these two motion vectors are averaged, even when they point to different reference images, and their reference image is defined as that of p0Cand; if only one motion vector is available, use it directly; if no motion vector is available, keep this list invalid. Furthermore, if the average interpolation filter indices of p0Cand and p1Cand are different, they are set to 0.

[00054] When the merge list is not complete after adding the average pairwise merge candidates, zero MVPs are inserted at the end until the maximum number of merge candidates is found. MERGER ESTIMATE REGION

[00055] The Merge Estimation Region (MER) allows independent derivation of the merge candidate list for CUs in the same merge estimation region (MER). A candidate block that is within the same MER as the current CU is not included in the generation of the current CU merge candidate list. Furthermore, the process of updating the candidate list for the history-based motion vector predictor is only updated if (xCb + cbWidth) >> Log2ParMrgLevel is greater than (xCb >> Log2ParMrgLevel) and (yCb + cbHeight) >> Log2ParMrgLevel is greater than (yCb >> Log2ParMrgLevel), where (xCb, yCb) is the position of the upper-left luminance sample of the current CU in the image and (cbWidth, cbHeight) is the size of the CU. The MER size is selected on the encoder side and signaled. Petition 870250094644, dated 10 / 16 / 2025, p. 27 / 111 20 / 94 as log2_parallel_merge_level_minus2 in the Sequence Parameter Set (SPS). Bi-prediction with weighting at the CU level (BCW)

[00056] In HEVC, the bi-prediction signal, Pbi-pred, is generated by averaging two prediction signals, PoE Plr, obtained from two different reference images and / or using two different motion vectors. In VVC, the bi-prediction mode is extended beyond simple averaging to allow weighted averaging of the two prediction signals. Pbi-pred = ((8 — w) * Po+ W * P±+ 4) » 3 (4) .

[00057] Five weightings are allowed in the biprediction weighted average, w E{-2,3,4,5,10}. For each bipredicted CU, the weighting w is determined in one of two ways: 1) for a non-merged CU, the weighting index is signaled after the motion vector difference; 2) for a merged CU, the weighting index is inferred from neighboring blocks based on the candidate merge index. BCW is applied only to CUs with 256 or more luminance samples (i.e., the CU width times the CU height is greater than or equal to 256). For low-latency images, all 5 weightings are used. For images without low latency, only 3 weightings (w E {3, 4, 5}) are used. In the encoder, fast search algorithms are applied to find the weighting index without significantly increasing the encoder complexity. These algorithms are summarized below. Details are disclosed in the VTM software and in document JVET-L0646 (Yu-Chi Su, et al.)., “CE4related: Generalized bi-prediction improvements combined from JVET-L0197 and JVET-L0296”, Joint Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29, 12th Meeting: Macau, CN, 3-12 October 2018, Document: JVET-L0646).

[00058] The BCW weighting index is encoded using a context-encoded file followed by deviation-encoded files. The first encoded file Petition 870250094644, dated 10 / 16 / 2025, page 28 / 111 21 / 94 by context indicates whether equal weighting is used; and if unequal weighting is used, additional files are flagged using offset coding to indicate which unequal weighting is used.

[00059] In VVC, CIIP and BCW cannot be applied together to a CU. When a CU is encoded with CIIP mode, the BCW index of the current CU is set to 2 (i.e., w=4 for equal weighting). Equal weighting implies the default value for the BCW index. Geometric Partitioning Mode (GPM)

[00060] In VVC, a geometric partitioning mode is supported for inter-prediction. The geometric partitioning mode is signaled using an indicator at the CU level as a merge mode type, with other merge modes including regular merge mode, MMVD mode, CIIP mode, and sub-block merge mode. In total, 64 partitions are supported by the geometric partitioning mode for each possible CU size wxft = 2mx2n with m,n G {3---6), excluding 8x64 and 64x8.

[00061] When this mode is used, a CU is divided into two parts by a geometrically located straight line (shown in Figure 10). The location of the dividing line is mathematically derived from the angle and displacement parameters of a specific partition. Each part of a geometric partition in the CU is interpredicted using its own motion; only single prediction is allowed for each partition, i.e., each part has a motion vector and a reference index. The single prediction motion constraint is applied to ensure that, as in conventional dual prediction, only two motion-compensated predictions are needed for each CU. The single prediction motion for each partition is derived.

[00062] If geometric partition mode is used for Petition 870250094644, dated 10 / 16 / 2025, page 29 / 111 22 / 94 is the current CU, then a geometric partition index indicating the partition mode of the geometric partition (angle and offset) and two merge indices (one for each partition) are additionally signaled. The maximum GPM candidate size is explicitly signaled in the SPS and specifies the binarization of the syntax for the GPM merge indices. After predicting each part of the geometric partition, the sample values ​​along the edge of the geometric partition are adjusted using a blending process with adaptive weightings. This is the prediction signal for the entire CU, and the transformation and quantization process will be applied to the entire CU, as in other prediction modes. Finally, the motion field of a CU predicted using the geometric partition modes is stored. BUILDING THE LIST OF CANDIDATES FOR THE SINGLE PREDICTION

[00063] The list of candidates for single prediction is derived directly from the list of candidates for fusion constructed according to the extended fusion prediction process. Let n denote the index of the single prediction motion in the geometric list of candidates for single prediction. The motion vector LX of the nth extended fusion candidate (X = 0 or 1, i.e., LX = L0 or L1), with X equal to the parity of n, is used as the nth single prediction motion vector for the geometric partitioning mode. If there is no corresponding motion vector LX of the nth extended fusion candidate, the motion vector L(1 - X) of the same candidate is used as the single prediction motion vector for the geometric partitioning mode. MIXING ALONG THE GEOMETRIC PARTITIONING EDGE

[00064] After predicting each part of a geometric partition using its own motion, the mixture is applied to the two prediction signals to derive samples around the edge of the geometric partition. The weighting of the mixture for each CU position is derived based on the distance between the individual positions. Petition 870250094644, dated 10 / 16 / 2025, page 30 / 111 23 / 94 and the edge of the partition. Combined Inter- and Intra-Prediction (CIIP)

[00065] In VVC, when a CU is encoded in merge mode, if the CU contains at least 64 luminance samples (i.e., the CU width times the CU height equals or is greater than 64) and if both the CU width and height are less than 128 luminance samples, an additional indicator is triggered to indicate whether the combined inter / intra prediction mode (CIIP) is applied to the current CU. As the name suggests, CIIP prediction combines an inter prediction signal with an intra prediction signal. The inter prediction signal in CIIP mode P_inter is derived using the same inter prediction process applied to regular merge mode; and the intra prediction signal P_intra is derived following the regular intra prediction process with planar mode.Next, the intra- and inter-prediction signals are combined using weighted averaging, where the weighting value wt is calculated depending on the encoding modes of the upper and left neighboring blocks (as shown in Figure 11) of the current CU 1110, as follows: - If the top neighbor is available and coded as intra, set isIntraTop to 1, otherwise set isIntraTop to 0; - If the left neighbor is available and intra-coded, set isIntraLeft to 1, otherwise set isIntraLeft to 0; - If (isIntraLeft + isIntraTop) equals 2, then wt is defined as 3; Otherwise, if (isIntraLeft + isIntraTop) equals 1, so wt is defined as 2; Otherwise, set wt to 1.

[00066] The CIIP prediction is formed as follows: Pciip = ((4 - wt) * Pinter+ wt* Pintra+ 2) » 2 (5). Multi-hypothesis prediction (MHP) (More details can be found) Petition 870250094644, dated 10 / 16 / 2025, page 31 / 111 24 / 94 (FOUND IN JVET-W2025)

[00067] In the inter-prediction mode with multiple hypotheses (JVET-M0425), one or more additional motion-compensated prediction signals are signaled, in addition to the conventional bi-prediction signal. The resulting overall prediction signal is obtained by sample-weighted superposition. With the bi-prediction signal p_bi and the first additional inter-prediction signal / hypothesis h_3, the resulting prediction signal p_3 is obtained as follows: p3= (1- a)pbi+ ah3.

[00068] The weighting factor α is specified by the new syntax element add_hyp_weight_idx, according to the mapping in Table 1. Table 1. Mapping between the weighting factor α and add_hyp_weight_idx. add_hyp_we i ght_i dx a 0 1 / 4 1 -1 / 8

[00069] Similarly to the above, more than one additional prediction signal can be used. The resulting global prediction signal is iteratively accumulated with each additional prediction signal. pn+l =(1 — an+l)pn+ an+lhn+l ·

[00070] The resulting global prediction signal is obtained as the last pn (i.e., the pn with the highest index n). For example, up to two additional prediction signals can be used (i.e., n is limited to 2).

[00071] The motion parameters of each additional prediction hypothesis can be signaled explicitly by specifying the reference index, the motion vector predictor index, and the motion vector difference, or implicitly by specifying a merge index. A separate multiple hypothesis merge indicator distinguishes Petition 870250094644, dated 10 / 16 / 2025, page 32 / 111 25 / 94 between these two signaling modes.

[00072] In the present invention, methods are disclosed for improving chroma prediction by deriving cross-component prediction using reference data, for example, offset prediction generated using motion information and / or block vectors or model reconstruction, of corresponding luma and chroma components. Furthermore, methods are disclosed for deriving inter-, intra-, or block vector prediction using regression-based derivation. BRIEF SUMMARY OF THE INVENTION

[00073] A method and apparatus for video encoding are disclosed. According to this method, input data associated with a current block in a current image comprising a first color component and a second color component are received, wherein the input data comprise pixel data to be encoded on the encoder side or data associated with the current block to be decoded on the decoder side, and wherein the current block comprises a first color block and a second color block. It is determined whether a target mode is applied to the current block.In response to applying the target mode to the current block: a target candidate cross-component predictor for the second color block is derived, wherein a target candidate cross-component model associated with the target candidate cross-component predictor is derived using reference data from the first corresponding color component and the second corresponding color component for the current block, wherein the reference data are associated with a reference region comprising a model of the current block or a predefined region indicated using a vector; and a final prediction is derived using the target candidate cross-component predictor. The second color block is encoded or decoded using the final prediction.

[00074] In one embodiment, the reference data Petition 870250094644, dated 10 / 16 / 2025, page 33 / 111 26 / 94 includes the compensated prediction of the first corresponding color component and the second corresponding color component for the current block. In one embodiment, when the vector is a block vector or the current block is encoded using the block vector, the compensated prediction of the first corresponding color component and the second corresponding color component for the current block is derived using block compensation according to the block vector. In another embodiment, when the vector is a motion vector or the current block is encoded using the motion vector, the compensated prediction of the first corresponding color component and the second corresponding color component for the current block is derived using motion compensation according to the motion vector.In one embodiment, the reference data is derived using the reconstruction of the first corresponding color component and the second corresponding color component for the reference region.

[00075] In one embodiment, the target candidate cross-component model is derived using current prediction samples, model samples, or both.

[00076] In one embodiment, after the target candidate cross-component model is derived, the target candidate cross-component predictor for the second color block is derived by applying the target candidate cross-component model to the reconstructed first color block.

[00077] In one embodiment, the target candidate cross-component model corresponds to the CCLM (Cross-Component Linear Model), MMLM (Multiple Model CCLM), or CCCM (Cross-Component Convolutional Model).

[00078] In one embodiment, the target candidate cross-component model is selected from a set of candidates comprising multiple candidates or referring to a list of candidates. In one embodiment, the set of Petition 870250094644, dated 10 / 16 / 2025, p. 34 / 111 27 / 94 candidates comprise one or more candidates for inherited cross-component models. In one embodiment, said one or more candidates for inherited cross-component models comprise one or more spatial candidates associated with one or more corresponding cross-component models from one or more neighboring blocks. In another embodiment, said one or more neighboring blocks correspond to one or more adjacent blocks or to one or more non-adjacent blocks. In one embodiment, said one or more candidates for inherited cross-component models comprise one or more temporal candidates associated with one or more corresponding cross-component models from one or more grouped temporal positions.In one embodiment, the aforementioned one or more candidate models between inherited components comprises one or more historical candidates associated with one or more models between corresponding components from a historical table that includes one or more models between corresponding components from previously coded blocks.

[00079] In one embodiment, (a) a candidate model associated with the target cross-component candidate model is selected from a set of self-derived candidates consisting of self-derived cross-component candidate models or (b) the candidate model associated with the target cross-component candidate model is selected from using a legacy model or using a self-derived model. In one embodiment, an index is explicitly signaled or parsed to select the candidate model in case (a), case (b), or a combination thereof. In one embodiment, the candidate model in case (a), case (b), or a combination thereof is implicitly selected. In one embodiment, the candidate model is implicitly selected using model-based mode derivation.

[00080] In one modality, a first syntax is signaled or parsed at a TU (Unit of) level. Petition 870250094644, dated 10 / 16 / 2025, page 35 / 111 28 / 94 Transformation), TB (Transformation Block) level, CU (Coding Unit) level, CB (Coding Block) level, or a combination thereof to indicate how to obtain the target cross-component candidate model for the current block or whether to apply the target mode to the current block. In one embodiment, the first syntax is signaled or parsed only when it satisfies that a corresponding TU or TB has a non-zero Cbf (coded block indicator equal to true), the current block is coded using a supported mode, or a size condition is met.

[00081] In one embodiment, when an IBC (Intra-Block Copy) related mode or an inter-prediction mode is used, the block vector-based prediction for the second color block or the inter-prediction for the second color block is combined with or replaced by the target candidate cross-component predictor.

[00082] In one embodiment, block vector-based prediction or inter-prediction is combined with the target candidate cross-component predictor using a weighting.

[00083] In one embodiment, when an inter-component prediction mode is used for the second color block, one or more hypotheses from one or more candidate cross-component models are combined with one or more hypotheses from the inter-component prediction mode.

[00084] In one embodiment, if a cross-component-related mode is used to generate prediction samples for the second color block and the current block is encoded using an inter- or IBC-related encoding tool, an indicator is flagged or analyzed to indicate whether the cross-component-related mode used is inherited from a previously encoded block or derived using a predetermined cross-component mode.

[00085] In one modality, the current block is divided into several sub-blocks. In one modality, each sub-block derives its own set of candidates or uses its own data from Petition 870250094644, dated 10 / 16 / 2025, page 36 / 111 29 / 94 reference. BRIEF DESCRIPTION OF THE DRAWINGS

[00086] Figure 1A illustrates an exemplary Inter / Intra adaptive video coding system that incorporates loop processing.

[00087] Figure 1B illustrates a corresponding decoder for the encoder in Figure 1A.

[00088] Figure 2A illustrates an example of a selected model for a current block, where the model comprises T rows above the current block and T columns to the left of the current block.

[00089] Figure 2B illustrates an example for T=3 and the HoGs (Gradient Histograms) are calculated for the pixels in the middle row and the pixels in the middle column.

[00090] Figure 2C illustrates an example of the amplitudes (ampl) for the intra-angular prediction modes.

[00091] Figure 3 illustrates an example of the mixing process, in which two intra-angular modes (M1 and M2) are selected according to the indices with the two highest bars of the histogram bars.

[00092] Figure 4 illustrates an example of intra-model-based derivation mode (TIMD), where TIMD implicitly derives the intra-prediction mode of a CU using a neighboring model in both the encoder and the decoder.

[00093] Figure 5 illustrates the 16 gradient patterns for the Gradient Linear Model (GLM).

[00094] Figure 6A illustrates the neighboring blocks used to derive space fusion candidates for VVC.

[00095] Figure 6B illustrates the possible candidate pairs considered for redundancy checking in VVC.

[00096] Figure 7 illustrates an exemplary pattern of non-adjacent spatial fusion candidates.

[00097] Figure 8 illustrates an example of derivation of Petition 870250094644, dated 10 / 16 / 2025, page 37 / 111 30 / 94 temporal candidate, in which a stepped motion vector is derived according to the POC (Picture Order Count) distances.

[00098] Figure 9 illustrates the position of the selected temporal candidate between candidates C0 and C1.

[00099] Figure 10 illustrates examples of GPM divisions grouped by identical angles. [000100] Figure 11 illustrates an example of the derivation of the weighting value for Combined Inter and Intra Prediction (CIIP) according to the coding modes of the upper and left neighboring blocks. [000101] Figure 12A illustrates the center of the pattern for selecting source terms for the luminance component. [000102] Figure 12B illustrates an example of a 5x5 cross pattern and Figure 12C illustrates an example of a 5x5 diamond pattern for the luminance component. [000103] Figure 13A illustrates the center of the pattern for selecting source terms for the chroma component. [000104] Figure 13B illustrates an example of a 5x5 cross pattern and Figure 13C illustrates an example of a 5x5 diamond pattern for the chroma component. [000105] Figure 14 illustrates an example of the spatial region adjacent to the current block, including the reference region above, the reference region to the left, and the reference region above to the left to derive the weighting configuration. [000106] Figure 15 illustrates an example of prediction derivation for a chroma sample using 6 non-subsampled source terms of the luma component. [000107] Figure 16 illustrates an example of reconstructed samples and predicted samples in the boundary region of a current block for boundary matching assessment. [000108] Figure 17 illustrates a flowchart of an exemplary video coding system that derives prediction between components using reference data from the components. Petition 870250094644, dated 10 / 16 / 2025, page 38 / 111 31 / 94 of corresponding luminance and chrominance, according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION [000109] It will be readily understood that the components of the present invention, as described and illustrated in the figures presented herein, can be arranged and designed in a wide variety of different configurations. Thus, the more detailed description that follows of 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 “an embodiment,” “in an embodiment,” or similar language mean that a particular feature, structure, or characteristic described in relation to the embodiment may be included in at least one embodiment of the present invention. Thus, occurrences of the phrases “in an embodiment” or “an embodiment” in various places throughout this specification do not necessarily all refer to the same embodiment. [000110] Furthermore, the described features, structures, or properties can be combined in any suitable manner in one or more embodiments. A person skilled in the relevant technical field 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 cases, 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 better understood with reference to the drawings, in which similar parts are designated by similar numbers throughout the document. The following description is intended only as an example and simply illustrates certain selected embodiments of apparatus and methods that are consistent with the invention herein. Petition 870250094644, dated 10 / 16 / 2025, p. 39 / 111 32 / 94 claimed. [000111] In this invention, methods for improving chroma prediction using cross-component prediction are disclosed. For example, compensated prediction generated using motion information and / or block vectors, or model reconstruction, of corresponding luma and chroma components is used to derive cross-component prediction. In addition, methods for deriving inter-, intra-, or block vector prediction using regression-based derivation are disclosed. I. Derivation Based on Inter-, Intra-, or Block Vector Prediction Regression [000112] In this invention, a new mechanism is proposed to improve the accuracy of the prediction. The prediction of the current block is formed by the combination of one or more proposed source terms and a proposed weighting configuration. As shown in expression (6) (or equation 6), pred(i, j) is a target (predicted) sample in the current block that can be obtained after this proposed mechanism, sourceTermSet0 includes one or more source terms of the luma component, sourceTermSet1 includes one or more source terms of the chroma components, and biasTermSet includes one or more bias terms. pred(i, j) = (sourceTermSet0(i, j) + sourceTermSet1(i, j) + ... + biasTermSet) with the proposed weighting context. (6) where (i, j) is a sample position in the current block. [000113] Equation (6) is just an example, and the proposed mechanism can use any subset or extension of sourceTermSet0, sourceTermSet1, and biasTermSet. Each sample or any subset of samples in the current block gets its target (predicted) sample according to equation (6). The contents of sourceTermSet0 are described in Section I.1, the contents of sourceTermSet1 in Section I.2, the contents of biasTermSet in Section I.3, and the derivation of the predictor using the proposed source terms and weighting configuration. Petition 870250094644, dated 10 / 16 / 2025, page 40 / 111 33 / 94 is described in Section I.4. Several encoding tools, including MHP and / or BCW and / or CIIP and / or chroma cross-component prediction mode fusion (e.g., TIMD-CCM), with our proposed mechanism, are presented in Section I.4. I.1 CONTENT OF SOURCETERMSET0(I, J) [000114] SourceTermSet0(i, j) includes one or more luminance source terms denoted as sourceTerm00, sourceTerm01, ... and / or sourceTerm0n-1. The value of n signifies the number of derivations for the source term set. In one embodiment, the source terms can be linear and / or non-linear terms, only linear terms and / or only non-linear terms. In another embodiment, n is a predefined value, such as 1, 2, ... or any positive integer. For example, the predefined value is fixed in the default. For another example, the default value is less than or equal to a maximum limit indicated by a syntax in the bitstream, where the syntax is at the block, CTU, CTB, slice, mosaic, image, SPS, PPS, image and / or sequence level. In another embodiment, n is determined according to the encoding information of the current block and / or sample position (i, j).For example, when the current block is encoded using a specific encoding tool, n is (1) fixed at a predefined value, (2) determined according to the block width, block height, block area, encoding information and / or sample information for the current block, (3) determined according to the encoding information and / or sample information for the adjacent or non-adjacent spatial reference region of the current block, and / or (4) determined according to the encoding information and / or sample information for the temporal reference region of the current block. In another embodiment, the pattern of the n derivations refers to a pattern defined as any subset of an M x N window region around or including the position (il, Jl). That is, (il, Jl) is used to derive the window and / or the pattern. Petition 870250094644, dated 10 / 16 / 2025, page 41 / 111 34 / 94 which means one or more positions to be used. For example, (il, Jl) refers to the center of the window and / or the pattern. However, (il, Jl) is not limited to referring to the center of the window and / or the pattern. If the target sample is luma, (il, Jl) is (i, J). If the target sample is chroma (cb or cr), (il, Jl) is the colocalized luma position of (i, j). [000115] For example, (il, Jl) refers to the center of the window and only the center (il, Jl) of the window is used, as illustrated in Figure 12A, where the center is indicated by a gray square C. On the other hand, (il, Jl) refers to the center of the window and the pattern is a 5x5 cross, which may or may not include the center in (il, Jl), as shown in Figure 12B, where the samples used as source terms are shown as squares filled with dots. For another example, (il, Jl) refers to the center of the window and the pattern is a 5x5 rhombus, which may or may not include the center (il, Jl), as shown in Figure 12C, where the samples used as source terms are shown as squares filled with dots. [000116] In another modality, different derivations refer to the origin terms of different prediction modes or different mode types. In one submodality, one or more derivations are of the intra-mode type, another or more derivations are of the inter-mode type, and / or another or more derivations are of the IBC mode type. In another submodality, one or more derivations are of the intra-MIP prediction modes, another or more derivations are of the intra-non-MIP prediction modes. [000117] For a source term in the set of source terms, the following methods are used to determine the generation of the source content. [000118] In one embodiment, the source content is based on a predicted sample generated by a prediction mode and / or on a reconstructed sample generated based on the sample predicted by a prediction mode and a reconstructed residual. For example, Petition 870250094644, dated 10 / 16 / 2025, page 42 / 111 35 / 94 Cross-component prediction generation for chroma using a model, the source content, which can be seen as the model input, such as the reconstruction from the corresponding luma, is combined or multiplied with the weighting, which can be seen as the model parameters, to derive the cross-component prediction for chroma. [000119] In a submodality, the prediction mode belongs to the intra mode type, the inter mode type, or a third mode type (e.g., IBC mode type). For example, if the prediction mode belongs to the intra mode type, the prediction mode refers to planar, DC, horizontal, vertical, other angular (directional) prediction mode, any intra prediction mode specified in the 67 / 131 intra prediction mode domain, wide-angle intra prediction modes (WAIP), TIMD-derived modes, DIMD-derived modes, intraTMP, and / or any intra prediction modes specified in the standard. For another example of the prediction mode belonging to the inter mode type, the prediction mode refers to jump mode, regular merge modes, MMVD modes, affine modes, SbTMVP, AMVR, any merge mode specified in the standard, any AMVP mode (advanced MVP where AMVP may be referred to as non-merged inter) specified in the standard, or any inter mode specified in the standard.For another example of a prediction mode belonging to the IBC mode type, the prediction mode refers to IBC fusion, IBC AMVP, or any IBC mode specified in the standard. It should be noted that any possible combination between the prediction mode and the mode type is supported in this invention. That is, any prediction mode mentioned may fall under any mode type according to the standard definition. For example, following the standard definition, if the IBC mode belongs to the inter mode type, the prediction mode belonging to the inter mode type in the embodiments may refer to an IBC mode. [000120] In another sub-modality, the source content is the Petition 870250094644, dated 10 / 16 / 2025, p. 43 / 111 36 / 94 filtered source or the source with any preprocessing. For example, the source content is the predicted or reconstructed sample after filtering with a predefined model or filter. For the example of generating cross-component prediction for chroma using a model, the source content, which can be seen as the model input, such as the reconstruction from the corresponding luma, is filtered using downscaling filters or gradient filters. [000121] In another sub-modality, the source content is the gradient information of the predicted and / or reconstructed samples. If the target sample (i, j) belongs to the chrominance and the gradient information of the colocalized luminance sample (as the central circle) is calculated with any of the following Sobel filters (as gradient filters shown in Figure 5) or any predefined filter. Each value around the central circle is multiplied by the corresponding predicted or reconstructed samples in the colocalized luminance block and then summed together to form the gradient information for the target sample source term (i, j). [000122] In another sub-modality, if the target sample belongs to a luma sample, the predicted sample and / or the reconstructed sample is located within the current block; otherwise (the target sample belongs to a chroma sample (cb or cr)), the predicted sample and / or the reconstructed sample is located within the colocalized block (luma) of the current block (chroma). The predicted sample and / or the reconstructed sample is treated as an initial sample and used as source content to generate the target sample. [000123] In another mode, the values ​​of the source terms are adjusted (added or subtracted) by a predefined deviation. If the target sample refers to luma, several modes are used to generate the deviation of the source term. Petition 870250094644, dated 10 / 16 / 2025, page 44 / 111 37 / 94 In one submodality, the deviation is determined as the average value of (or any subset of) predicted or reconstructed samples in the current block or the reference region of the current block. In another submodality, the deviation is determined as a sample value from predefined predicted or reconstructed samples in the current block or the reference region of the current block. For example, the sample value is from the upper left position (just outside the upper left corner of the current block). If the target sample refers to chrominance, several modalities are used to generate the deviation of the source term. In one submodality, the deviation is determined as the average value of (or any subset of) predicted or reconstructed samples in the luminance block colocalized from the current block (chrominance) or in the reference region of the colocalized luminance block.In another sub-mode, the offset is determined as a sample value from predefined predicted or reconstructed samples in the colocalized luma block or the reference region of the colocalized luma block. For example, the sample value is from the top left position (just outside the top left corner of the colocalized luma block). [000124] In another embodiment, the source term may also include location information. For example, if the target sample refers to luminance, the horizontal location (i) of (i, j) is used in a source term and the vertical location (j) of (i, j) is used in a source term; otherwise, the horizontal location of the colocalized luminance block of sample (i, j) is used in a source term and the vertical location of the colocalized luminance block of sample (i, j) is used in a source term. I. 2 CONTENT OF SOURCETERMSET1(I, J) [000125] SourceTermSet1(i, j) includes one or more chroma source terms (cb or cr) denoted as sourceTerm10, sourceTermll, ... and / or sourceTerm1m-1. The value of m means the Petition 870250094644, dated 10 / 16 / 2025, p. 45 / 111 38 / 94 number of derivations for the set of source terms. In one embodiment, the source terms can be linear and / or non-linear terms, only linear terms and / or only non-linear terms. In another embodiment, m is a predefined value, such as 1, 2, ... or any positive integer. For example, the predefined value is fixed in the pattern. For another example, the default value is less than or equal to a maximum limit indicated by a syntax in the bitstream, where the syntax is at the block, CTU, CTB, slice, mosaic, image, SPS, PPS, picture and / or sequence level. In another embodiment, m is determined by the encoding information of the current block and / or the sample position (i, j).For example, when the current block is encoded by a specific encoding tool, m is (1) fixed at a predefined value, (2) determined according to the block width, block height, block area, encoding information and / or sample information for the current block, (3) determined according to the encoding information and / or sample information for the adjacent or non-adjacent spatial reference region of the current block, and / or (4) determined according to the encoding information and / or sample information for the temporal reference region of the current block. In another embodiment, the pattern of the m derivations refers to a pattern defined as any subset of an M2 x N2 window region around or including the position (ic, jc). That is, (ic, jc) is used to derive the window and / or the pattern, meaning one or more positions to use. In some examples, (ic, jc) refers to the center of the window and / or the pattern.However, (ic, jc) is not limited to referring to the center of the window and / or pattern. If the target sample is for chroma (cb or cr), (ic, jc) is (i, j). If the target sample is for luma, (ic, jc) is the chroma position colocalized from (i, j). [000126] For example, (ic, jc) refers to the center of the window and only the center (ic, jc) of the window is used, as Petition 870250094644, dated 10 / 16 / 2025, p. 46 / 111 39 / 94 illustrated in Figure 13A, where the center is indicated by a gray square C. On the other hand, (ic, jc) refers to the center of the window and the pattern is a 5x5 cross, which may or may not include the center in (ic, jc), as shown in Figure 13B, where the samples used as source terms are shown as squares filled with dots. For another example, (ic, jc) refers to the center of the window and the pattern is a 5x5 rhombus, which may or may not include the center (ic, jc), as shown in Figure 13C, where the samples used as source terms are shown as squares filled with dots. [000127] In another modality, different derivations refer to the origin terms of different prediction modes or different mode types. In one submodality, one or more derivations are of the intra-mode type, another or more derivations are of the inter-mode type, and / or another or more derivations are of the IBC mode type. In another submodality, one or more derivations are of the intra-MIP prediction modes, another or more derivations are of the intra-non-MIP prediction modes. [000128] For a source term in the set of source terms, the following methods are used to determine the generation of the source content. [000129] In one embodiment, the source content is based on a predicted sample generated by a prediction mode and / or on a reconstructed sample generated based on the predicted sample by a prediction mode and a reconstructed residual. [000130] In a submodality, the prediction mode belongs to the intra mode type, the inter mode type, or a third mode type (e.g., IBC mode type). For an example of a prediction mode belonging to the intra mode type, the prediction mode refers to planar, DC, horizontal, vertical, other angular (directional) prediction modes, any intra prediction modes specified in the intra prediction mode domain 67 / 131, wide-angle intra prediction modes (WAIP), modes Petition 870250094644, dated 10 / 16 / 2025, p. 47 / 111 40 / 94 derived from TIMD, modes derived from DIMD, intraTMP, DBV, any of the cross-component modes (CCLM (including CCLM_LT, CCLM_L and / or CCLM_T), MMLM (including MMLM_LT, MMLM_L and / or MMLM_T), CCCM (including CCCM_LT, CCCM_L and / or CCCM_T), GLM and / or any variation or extension of the above modes) and / or any intra prediction modes specified in the standard. For another example of a prediction mode belonging to the inter mode type, the prediction mode refers to the skip mode, regular fusion modes, MMVD modes, affine modes, SbTMVP, AMVR, any fusion mode specified in the standard, any AMVP mode specified in the standard, or any inter mode specified in the standard. For another example of a prediction mode belonging to the IBC mode type, the prediction mode refers to IBC fusion, IBC AMVP (advanced MVP, where AMVP may be referred to as inter-non-fusion in this disclosure), or any IBC mode specified in the standard.Note that any possible combination between the prediction mode and the mode type is supported in this invention. That is, any prediction mode mentioned can be under any mode type according to the standard definition. For example, following the standard definition, if the IBC mode belongs to the inter mode type, the prediction mode belonging to the inter mode type in the embodiments can refer to an IBC mode. In one embodiment, DBV can be seen as using IBC to generate predicted chroma samples. [000131] In another sub-modality, the source content is the filtered source or the source with any pre-processing. For example, the source content is the predicted or reconstructed sample after filtering with a predefined model or filter. [000132] In another submodality, the source content is the gradient information of the predicted and / or reconstructed samples. If the target sample (i, j) belongs to the luma and the gradient information of the colocalized chroma sample is Petition 870250094644, dated 10 / 16 / 2025, page 48 / 111 41 / 94 calculated using any of the Sobel filters, any of the gradient filters, or any predefined filter. [000133] In another sub-modality, if the target sample belongs to a chroma sample, the predicted sample and / or the reconstructed sample is located within the current block; otherwise (the target sample belongs to a luma sample), the predicted sample and / or the reconstructed sample is located within the colocalized block (chroma) of the current block (luma). The predicted sample and / or the reconstructed sample is treated as an initial sample and used as source content to generate the target sample. [000134] In another mode, the values ​​of the source terms are adjusted (added to or subtracted from) by a predefined deviation. If the target sample refers to chrominance, several modes are used to generate the deviation of the source term. In one submode, the deviation is determined as the average value of the predicted or reconstructed samples (or any subset thereof) in the current block or the reference region of the current block. In another submode, the deviation is determined as a sample value from predefined predicted or reconstructed samples in the current block or the reference region of the current block. For example, the sample value is from the upper left position (just outside the upper left corner of the current block). If the target sample refers to luminance, several modes are used to generate the deviation of the source term.In one submodality, the offset is determined as the average value of the predicted or reconstructed samples (or any subset thereof) in the colocalized chroma block from the current block (luma) or in the reference region of the colocalized chroma block. In another submodality, the offset is determined as a sample value from predefined predicted or reconstructed samples in the colocalized chroma block or in the reference region of the colocalized chroma block. For example, the value... Petition 870250094644, dated 10 / 16 / 2025, page 49 / 111 42 / 94 of the sample is from the upper left position (just outside the upper left corner of the colocalized chroma block). [000135] In another embodiment, the source term may also include location information. For example, if the target sample refers to chrominance, the horizontal location (i) of (i, j) is used in a source term and the vertical location (j) of (i, j) is used in a source term; otherwise, the horizontal location of the colocalized chroma block of sample (i, j) is used in a source term and the vertical location of the colocalized chroma block of sample (i, j) is used in a source term. I.3 CONTENTS OF THE BIASTERMSET [000136] The bias term is any predefined value. In one embodiment, the bias term is a midValue according to the bitDepth specified in the standard. For example, the bias term is defined as (1 ⋅ (bitDepth-1)). In another embodiment, the bias term is the same for each sample in the current block. That is, the bias term is independent of the position (i, j). I.4 Derivation of the Predictor for Sample (I, J) I.4.1. PROPOSED WEIGHTING CONFIGURATION [000137] The proposed weighting configuration consists of estimating the relationship (e.g., minimizing distortion) between the combined results of these source terms and the reconstructed samples in the current block's reference region by a predefined regression method, to generate a weighting (referring to the model parameters) according to the regression method, and then applying the weighting to the source terms to obtain the target (predicted) samples in the current block. In one embodiment, the predefined regression method may be the linear minimum mean squared error (LMMSE) method as cross-component modes, e.g., CCLM, or it may be any method unified with the regression method used for modes of Petition 870250094644, dated 10 / 16 / 2025, page 50 / 111 43 / 94 cross-components, for example, CCLM. In another embodiment, the predefined regression method may be the LDL decomposition method as CCCM or it may be any method unified with the regression method used for CCCM. In another embodiment, the predefined regression method may be Gaussian elimination. For example, the chroma component reference data are derived using the reconstructed samples in the chroma block reference region as the estimation target, and the corresponding luma component reference data are derived using the reconstructed samples in the corresponding luma block reference region as source terms. Then, the target cross-component model is derived using the chroma and luma component reference data. [000138] In one embodiment, the reference region of the current block is the spatially adjacent or non-adjacent neighboring region of the current block 1410, as shown in Figure 14. The spatial neighboring region of the current block (as a model of the current block) includes the upper reference region 1420, the left reference region 1430, the upper left reference region 1440, and / or the default regression method may be the LDL decomposition method as CCCM or may be any method unified with the regression method used for CCCM. In another embodiment, the default regression method may be Gaussian elimination. For example, the chroma component reference data are derived using the reconstructed samples in the chroma block reference region as the estimation target, and the corresponding luma component reference data are derived using the reconstructed samples in the corresponding luma block reference region as source terms.Next, the target cross-component model is derived using the reference data for the chroma and luma components. [000139] In one embodiment, the reference region of the current block is the spatially adjacent or non-adjacent neighboring region. Petition 870250094644, dated 10 / 16 / 2025, page 51 / 111 44 / 94 of the current block 1410, as shown in Figure 14. The spatial [000140] The neighboring region of the current block (as a model of the current block) includes the upper reference region 1420, the left reference region 1430, the upper left reference region 1440 and / or any subset of the foregoing. The size of the upper reference region is Aw x AH, the size of the left reference region is Lw x LH and the size of the upper left reference region is ALW x ALH, where: - Aw = current block width (W), k*W, W + current block height (H), any predefined value or any adaptive value depending on the block position, block width, block height and / or current block area. - AH or ALH = H, any predefined value (1, 2, 4, ...), or any adaptive value depending on the block position, block width, block height, and / or current block area. - Lw or ALW = W, any predefined value (1, 2, 4, ...), or any adaptive value depending on the block position, block width, block height, and / or current block area. - LH = H, k*H, H + W, any predefined value, or any adaptive value depending on the block position, block width, block height, and / or current block area. I.4.2. DIFFERENT EXEMPLARY EXPRESSIONS I.4.2.1 predc(i,j) = a0· G(i,j) + at· rec'L(i,j) + a2· bias [000141] In this expression of the target sample being chroma, sourceTermSet0 includes two derivations as G(i, j) and rec'L(i,j), sourceTermSet1 is not used, and biasTerm refers to another derivation as midValue. G(i,j) is the gradient information generated from a selected gradient filter, and rec'L(i,j) is the luminance sample reconstructed with resolution reduction. The model parameters (a0, a1, and a2) of the weighting are derived based on: Using six adjacent sample rows and columns as the reference region for the current block. - Using the LDL decomposition method as a method of Petition 870250094644, dated 10 / 16 / 2025, page 52 / 111 45 / 94 regression I.4.2.2 . predc(i,j) = a0· C + al· Gy(i,j) + a2· Gx(i,f) + a3· Y(i,j) + a4· X(i,j) + a5· P(i,j) + a6· bias [000142] In this expression (similar to JVET-AC0054) of the target sample being chroma, sourceTermSet0 includes six touches as C (the reconstructed sample of colocalized / corresponding luma), Gy(i, j), Gx(i, j), Y, X and P (e.g., a non-linear term like CCCM), sourceTermSet1 is not used and biasTerm refers to another sample as midValue. - Gy (i,j) is the gradient information generated from a vertical gradient filter. - Gx (i,j) is the gradient information generated from a horizontal gradient filter. - Y and X are the vertical and horizontal locations of the colocalized luminance sample. - Using six adjacent rows and columns of samples as the reference region for the current block. - Using the LDL decomposition method as a regression method. I.4.2.3 predL(i,j) = a0· Pmode_0(i, / ) + al· Pmode_l(i, / ) + —+ as_l· Pmode_s — l(i, / )+ as· bias [000143] In this expression of the target sample being luma for an inter-encoded block, sourceTermSet0 includes s derivations such as Pmode_0 to Pmode_s-1, sourceTermSet1 is not used and biasTerm refers to another derivation such as midValue. Each or any subset of Pmode_0(i, j) to Pmode_s1(i, j) is the predicted sample from the mode indicated by an intermode index. For example, Pmode_0(i, j) is the predicted sample from the first mode derived using the intermode index. For example, Pmode_s-1(i, j) is the predicted sample of the s-th mode (or (s-1)-th mode, if the first mode is the 0th mode) derived using the intermode index. In a mode, a list of inter-mode candidates is first constructed. Petition 870250094644, dated 10 / 16 / 2025, page 53 / 111 46 / 94 inter-candidate lists are used in the inter-candidate list. For example, the inter-candidate list refers to a list of merge or AMVP candidates that is the same as or different from the list of merge or AMVP candidates for the normal merge or AMVP mode. For example, the merge candidate list for the normal merge mode is reused to derive the merge candidate list for the proposed mechanism. In another embodiment, only single prediction candidates, only double prediction candidates, or single and / or double prediction candidates are included in the inter-candidate list. In another embodiment, the maximum number of candidates in the inter-candidate list is specified in the standard as a fixed number or as a syntax such as block-level, CTU-level, SPS-level, PPS-level, slice-level, mosaic-level, image-level, and / or sequence-level flags.In another embodiment, an inter-candidate mode index is signaled or parsed to indicate each mode (mode 0 as or s-1). In another embodiment, an inter-candidate mode index is signaled or parsed to indicate a mode (e.g., mode 0), and for the remaining modes, these are selected according to mode 0. In another embodiment, two candidate lists are created. The mode or modes can be selected from one of the two candidate lists or from both candidate lists. One list is an inter-candidate list containing one or more inter-candidate candidates, and the other is an intra-candidate list containing one or more intra-candidate candidates. If a list contains only one candidate, it is assumed that this single candidate will be used without signaling. The parameters a0 to as-1 can be seen as a weighting for combining each predictor of the mode to be combined. [000144] In another embodiment, the proposed mechanism is treated as an optional mode (for example, an optional mode of MHP). That is, an indicator is signaled or analyzed in Petition 870250094644, dated 10 / 16 / 2025, page 54 / 111 47 / 94 encoder or decoder to indicate whether to use the proposed mechanism for the current block (MHP encoded). In one sub-mode, the indicator is at the block level, CTU level, slice level, SPS level, mosaic level, PPS level, and / or image level. In another mode, the indicator is context-encoded. For example, only one context is used to signal the indicator. For another example, the selection of the indicator context depends on the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. [000145] In another embodiment, the proposed mechanism is a replacement method. When generating the predictors of the current MHP encoded block with support from the proposed mechanism, the predictor generation is inferred to follow the proposed mechanisms. [000146] In another embodiment, here, s is a predefined value. For example, s = 1, 2, 3, or any positive integer. In a subemphasis, s is fixed at the predefined value in the standard. In another subemphasis, s is adaptable according to the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. For example, if the block width, height, or area is greater than a predefined limit, s is a larger number; otherwise, s is a smaller number. [000147] In another embodiment, when using the regression method to derive the model parameters in the current block reference region, the distortion to be minimized is between the resulting combination, including (1) predictors generated from mode 0 to s-1 in the current block reference region and (2) bias and (3) weighting configuration, and the samples reconstructed in the current block reference region. [000148] In another mode, sourceTermSet1 can be Petition 870250094644, dated 10 / 16 / 2025, page 55 / 111 48 / 94 is used in the expression. That is, the corresponding chroma information can be used to generate the target luma samples. [000149] In another embodiment, the MHP expression can be replaced by BCW, GPM, CIIP, and / or any luma or chroma coding tools that use multiple prediction hypotheses to form the final prediction of the current block, to apply the proposed mechanism when the current block uses the given coding tool. When the proposed mechanism is applied to BCW, each prediction hypothesis refers to a unique prediction of different lists (list0 or list1), and the expression for generating the final prediction is shown below. The prediction hypotheses for list0 and list1 are indicated with an inter-shared mode index, such as a candidate merge index or a candidate AMVP index. predL(i,j) = a0· Plist_0(i, / ) + al· Plist_1(i, / ) + a2· bias . I.4.2.4 predL(i,j) = a0· Ppat_0(i, / ) + al· Ppat_l(i, / ) + —+ as_l· Ppat_s — l(i, / )+ as· bias [000150] In this expression of the target sample being luma for an inter-encoded block, sourceTermSet0 includes s derivations such as Ppat_0 to Ppat_s-1, sourceTermSetl is not used, and biasTerm refers to another derivation such as midValue. Each or any subset of Ppat_0(i, j) to Ppat_s1(i, j) is the combined predicted sample following the inter-coding tool rule for the current block. For an example of the inter-coding tool being MHP, the combined predicted sample is formed by an MHP weighted average of each predictor generated from an MHP prediction mode to be combined. 0 to s-1 indicates the s-derivation pattern (any pattern predefined in Section I) of the source terms. [000151] In one embodiment, the proposed mechanism is treated as an optional mode (e.g., an optional mode of MHP). That is, an indicator is signaled or analyzed in the encoder or Petition 870250094644, dated 10 / 16 / 2025, page 56 / 111 49 / 94 decoder to indicate whether the proposed mechanism should be used for the current block (MHP encoded). In one sub-mode, the indicator is at the block level, CTU level, slice level, SPS level, block level, PPS level, and / or image level. In another mode, the indicator is context-encoded. For example, only one context is used to signal the indicator. For another example, the selection of the indicator context depends on the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. [000152] In another embodiment, the proposed mechanism is a replacement method. When generating the predictors of the current MHP encoded block with the support of the proposed mechanism, the predictor generation is inferred to follow the proposed mechanisms. [000153] In another embodiment, here, s is a predefined value. For example, s = 1, 2, 3, or any positive integer. In a subemphasis, s is fixed at the predefined value in the standard. In another subemphasis, s is adaptable according to the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. For example, if the block width, height, or area is greater than a predefined limit, s is a larger number; otherwise, s is a smaller number. [000154] In another embodiment, when using the regression method to derive the model parameters in the current block reference region, the distortion to be minimized is between the resulting combination, including (1) the combination of predictors generated from the 0 to s-1 derivation pattern in the current block reference region and (2) the bias and (3) the weighting configuration, and the reconstructed samples in Petition 870250094644, dated 10 / 16 / 2025, page 57 / 111 50 / 94 current block reference region. [000155] In another embodiment, sourceTermSet1 can be used in the expression. That is, the corresponding chroma information can be used to generate the target luma samples. [000156] In another embodiment, the expression MHP can be replaced by BCW, GPM, CIIP and / or any luminance encoding tools to apply the proposed mechanism when the current block uses the given encoding tool. When the given encoding tool is GPM, generating combined predictors in the reference region of the current block or within the current block, the GPM weighted average follows the GPM partitioning line.In other words, following the GPM mixing rule, for samples close to the partitioning line, equal weighting is used for both predictions of the GPM prediction modes to be mixed; for samples distant from the partitioning line, a higher weighting is used for the prediction of one of the GPM modes to be combined and a lower weighting is used for the other prediction of the other GPM mode to be combined, if the current sample is located in the prediction unit belonging to one of the GPM modes to be combined (not belonging to the other GPM mode to be combined). I.4.2.5 predc(i,j) = a0· P_CCM_0(i, / ) + —+ as_±· P_CCM_s_i (ij) + asbias [000157] In this expression of the target sample being chroma for a cross-component mode (CCM) encoded block, sourceTermSet1 includes s derivations such as P_CCM_0 to P_CCM_s-1, sourceTermSet0 is not used, and biasTerm refers to another derivation such as midValue. Each or any subset of P_CCM_0(i, j) to P_CCM_s1(i, j) is the predicted sample of the selected mode from all or any subset of the component prediction modes. Petition 870250094644, dated 10 / 16 / 2025, p. 58 / 111 51 / 94 candidates cross-referenced for the encoding mode. For example, P_CCM_0(i, j) is the predicted sample of the first mode. For example, P_CCM_s-1(i, j) is the predicted sample of the s-th mode (or (s1)-th mode if the first mode is the 0-th mode). [000158] In a mode, s is defined as at least two. For the example of s being 2, when the encoding mode is CCM, one predictor of CCM_0 and the other predictor of CCM_1 are used to generate the final predictor. In a submode, CCM_0 is indicated by a mode index and CCM_1 is derived by the TIMD derivation process. That is, CCM_0 is selected depending on the signaled chroma prediction mode index and CCM_1 is the mode that has the lowest TIMD cost (among a predefined set of candidate CCMs) in the reference region (i.e., model) of the current block. For example, the predefined set of candidate CCMs for CCM includes MMLM_L, MMLM_T, and / or MMLM_LT. For another example, CCM_0 is one of the following: CCLM_L, CCLM_T, and CCLM_LT. [000159] In another embodiment, the proposed mechanism is treated as an optional mode (e.g., an optional mode of the CCM-encoded block). That is, an indicator is signaled or analyzed in the encoder or decoder to indicate whether the proposed mechanism should be used for the current block. In a sub-embodiment, the indicator is at the block level, CTU level, slice level, SPS level, mosaic level, PPS level, and / or image level. In another embodiment, the indicator is context-encoded. For example, only one context is used to signal the indicator. For another example, the indicator context selection depends on the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. [000160] In another embodiment, the proposed mechanism is a substitution method. When generating the block predictors Petition 870250094644, dated 10 / 16 / 2025, p. 59 / 111 Given the current 52 / 94 encoded CCM with support for the proposed mechanism, predictor generation is inferred to follow the proposed mechanisms. [000161] In another embodiment, here, s is a predefined value. For example, s = 1, 2, 3, or any positive integer. In a subemphasis, s is fixed at the predefined value in the standard. In another subemphasis, s is adaptable according to the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. For example, if the block width, height, or area is greater than a predefined limit, s is a larger number; otherwise, s is a smaller number. [000162] In another embodiment, when using the regression method to derive the model parameters in the current block reference region, the distortion to be minimized is between the resulting combination, including (1) predictors generated from mode 0 to s-1 in the current block reference region, (2) bias and (3) weighting configuration, and the samples reconstructed in the current block reference region. [000163] In another embodiment, sourceTermSet0 can be used in the expression. That is, the corresponding luminance information can be used to generate the target chrominance samples. For example, rec'L(i,j), G(i,y), and / or Gy(i,j), Gx(i,j) are added as source terms in sourceTermSet0. I.4.2.6 . predc(i,j) = a0· Ppat_0(i, / ) + al· Ppat_l(i, / ) + — + as_l· Ppat_s — l(i, / )+ as· bias [000164] In this expression of the target sample being chroma for a CCM encoded block, sourceTermSet1 includes s derivations such as Ppat_0 to Ppat_s-1, sourceTermSet0 is not used and biasTerm refers to another derivation such as midValue. Each or any subset of Ppat_0(i, j) to Ppat_s1(i, j) is the predicted combined sample following the CCM coding tool rule for the current block. For example Petition 870250094644, dated 10 / 16 / 2025, pp. 60 / 111 53 / 94 of the CCM coding tool being TIMD-CCM, the predicted combined sample is formed by a TIMD-CCM weighted average of each predictor generated from a TIMD-CCM prediction mode to be combined. 0 to s-1 indicates the s-derivation pattern (any pattern predefined in Section I) of the source terms. An example of a TIMD-CCM weighted average is shown below. - In one case, both of the two CCMs to be combined are selected by one or more signed mode indices. In another case, only one of the two CCMs to be combined is selected by a signed mode index, and the other of the two CCMs to be combined is determined by the TIMD derivation process. The TIMD derivation process means that, for each of all candidate CCMs for the other of the two CCMs to be combined, a TIMD cost (or model cost) is calculated in the model by comparing the skewness between the samples reconstructed in the model and the samples predicted in the model, and the candidate CCM with the lowest TIMD cost is determined as the other of the two CCMs to be combined, where the samples predicted in the model are generated by: - The current candidate CCM models are derived with the inputs as (1) the model reference region for the current chroma block and (2) the model reference region for the colocalized luma block. The models derived from the current candidate CCM would be applied to the reconstructed samples in the colocalized luma block model to obtain the predicted samples in the current block (chroma) model. - In another case, the two CCMs to be combined are determined by the TIMD derivation process. After deciding which two CCMs to combine, the weighting for the two CCMs to be combined depends on the TIMD costs of the two CCMs to be combined. The mode with a lower TIMD cost gets a higher weighting when performing the TIMD-CCM weighted average (or so-called model-based CCM). Petition 870250094644, dated 10 / 16 / 2025, pp. 61 / 111 54 / 94 [000165] In one embodiment, the proposed mechanism is treated as an optional mode (e.g., an optional mode of TIMDCCM). That is, an indicator is signaled or analyzed in the encoder or decoder to indicate whether the proposed mechanism should be used for the current block (encoded by TIMDCCM). In a sub-emphasis, the indicator is at the block level, CTU level, slice level, SPS level, block level, PPS level, and / or image level. In another embodiment, the indicator is context-encoded. For example, only one context is used to signal the indicator. For another example, the indicator context selection depends on the encoding information, block width, block height, and / or block area of ​​the current block and / or the encoding information, block width, block height, and / or block area of ​​the neighboring block. [000166] In another embodiment, the proposed mechanism is a replacement method. When generating the predictors of the current block encoded by TIMD-CCM with the support of the proposed mechanism, the predictor generation is inferred to follow the proposed mechanisms. [000167] In another embodiment, here, s is a predefined value. For example, s = 1, 2, 3, or any positive integer. In a subemphasis, s is fixed at the predefined value in the standard. In another subemphasis, s is adaptable according to the coding information, block width, block height, and / or block area of ​​the current block and / or the coding information, block width, block height, and / or block area of ​​the neighboring block. For example, if the block width, height, or area is greater than a predefined limit, s is a larger number; otherwise, s is a smaller number. [000168] In another embodiment, when using the regression method to derive the model parameters in the reference region of the current block, the distortion to be minimized is between the resulting combination, including (1) the combination of predictors Petition 870250094644, dated 10 / 16 / 2025, pp. 62 / 111 55 / 94 generated from the standard derivation 0 to s-1 in the current block reference region, (2) the polarization and (3) the weighting configuration, and the samples reconstructed in the current block reference region. [000169] In another embodiment, sourceTermSet0 can be used in the expression. That is, the corresponding luminance information can be used to generate the target chrominance samples. [000170] In another embodiment, the expression TIMD-CCM can be replaced by any cross-component tools, which use multiple prediction hypotheses from multiple CCM models to form the final prediction, to apply the proposed mechanism when the current block uses a given coding tool. [000171] In another embodiment, when generating the target predictors of the current block and / or generating the model predictors in the reference region of the current block, a long-touch post-filter is applied. The filtering method can be any pattern proposed in the invention above. II. Chroma Prediction Through Component Information and / or Fusion [000172] Intercomponent information is used to improve the prediction accuracy of an interblock. To improve the prediction accuracy of the chroma component of the interblock, luma information from the corresponding luma component and / or chroma information from the previously encoded chroma component is used. The first scheme is for a single coding unit (under single tree division) including luminance (Y) and chrominance (Cb and / or Cr) components, where the prediction for Cb and / or Cr is improved using Y information. The second scheme consists of a coding unit (under single tree division) including components Petition 870250094644, dated 10 / 16 / 2025, pp. 63 / 111 56 / 94 of luminance (Y) and chrominance (Cb and / or Cr) or for a coding unit (under double chrominance tree splitting) including chrominance components (Cb and / or Cr), the prediction for Cr is improved using Cb information. For example, derive model parameters using reconstructed neighboring samples of Cb and Cr as X inputs, which are used to predict Cr when deriving the model for between-component prediction, and Y, which is used as the target when deriving the model for between-component prediction, from the model derivation. Then, generate the Cr prediction from the derived model parameters and the reconstructed Cb samples. [000173] Next, several modalities related to the first scheme are proposed for (1) determining how to obtain one or more model information that may come from a legacy method in Section II.1, determining how to obtain one or more model information that may come from any derivation method (e.g., the regression-based method in Section I and / or Section II.2), and / or constructing a candidate set (as in Section II.1, e.g., the set referring to the list) for the current block, where the candidate set or list includes cross-component models, (2) deriving one or more model information or selecting one or more model information from the set (e.g., the set referring to the list) as per Section II.3 and / or (3) as per Section II.4. Use the model information (similar to and / or unified with the intrachroma cross-component model or mode) to generate one or more predictive hypotheses for the current chroma component (Cb or Cr) by applying and / or modifying the selected model information to the reconstructed or predicted samples for the corresponding luminance component. When the selected model information refers to traditional linear models between components, the proposed method is called model mode. Petition 870250094644, dated 10 / 16 / 2025, pp. 64 / 111 57 / 94 linear between components (inter CCLM). When the selected model information refers to convolutional models between components derived by a regression-based method (such as CCCM and / or any methods proposed in Section I and / or Section II.2, for example), the proposed method is called a convolutional model mode between components (inter CCCM). Details can be found in Section II.5. [000174] The proposed modalities can also be used for the second scheme, using the previously coded chromatic component (Cb) as the luminance component in the first scheme. II.1 CONSTRUCT A SET OR LIST OF CANDIDATES INCLUDING MODELS AMONG COMPONENTS [000175] In one embodiment, a candidate set is constructed comprising multiple candidates or referring to a list of candidates. When constructing the set or list of candidate models similar to a merge (where the set or list is represented as modelList), one or more of the following candidate model information, e.g., inherited candidates, is included. - Information from the spatial model of spatial neighboring blocks (corresponding to “MVP spatial of spatial neighboring CUs for inter). - Information from the temporal model of colocalized blocks (corresponding to “temporal MVP of colocalized CUs for inter). - Model information based on the history of a FIFO table (corresponding to "MVP based on the history of a FIFO table for inter). - Average pairwise model information (corresponding to "average pairwise MVP for inter). - Standard model information (corresponding to “Zero MVs for inter). Petition 870250094644, dated 10 / 16 / 2025, pp. 65 / 111 58 / 94 [000176] In a candidate-type submodality being “Spatial model information of spatial neighboring blocks, one or more valid spatial neighboring blocks may be from one of the adjacent and / or non-adjacent spatial neighbors (or any subset of the blocks in a neighboring search region for the current block) that satisfies a predefined condition.For example, the predefined condition corresponds to the case where the neighbor is encoded by a cross-component mode (such as CCLM, MMLM, CCCM, GLM, the mode with mode or model information inherited from a fusion-like candidate list, multiple hypothesis (MH) CCLM that combines multiple cross-component prediction hypotheses to derive the final prediction for an MH CCLM-encoded block and / or any cross-component mode with syntax that does not belong to the traditional intra-component prediction modes) or combining with the cross-component mode (such as chroma fusion (or so-called LM-assisted Angular or Planar Mode), inter-CCLM or any variations specified in section II.5 (e.g., inter-CCCM) and / or any traditional mode with syntax that does not belong to the cross-component modes but uses the cross-component information to generate the prediction).When scanning neighboring space blocks, a candidate is added to the list if it is valid. The following shows some scanning orders when adding the space model information of neighboring space blocks (as shown in Figure 6A) to the list. For example, the scanning order follows B0 (above) A0 (left) -^ B1 (above right) -^ A1 (below left) -^ B2 (above left) or any predefined order. For example, the scanning order follows B1 -> A1 ^ B0 ^ A0. For example, the scanning order follows adjacent candidates before (after) non-adjacent candidates. [000177] In another submodality of the temporal model of colocalized blocks, the colocalized block is of Petition 870250094644, dated 10 / 16 / 2025, pp. 66 / 111 59 / 94 block in the reference image or colocalized as inter mode. For example, when the current block is encoded by the inter prediction mode, the colocalized block is referenced by the motion information (including motion vectors and / or the reference image) of the current block. If the current block is a sub-block motion mode (e.g., affine mode), each sub-block in the current block has its own colocalized temporal model information and / or any or all subsets of the colocalized temporal model information referenced by the motions of different sub-blocks are added to the list. For another example, the temporal model information might be from the colocalized block referenced by the motion information of neighboring blocks to the current block.If the proposed methods are applied to an IBC block or any mode that uses block vectors, the block vector information is used as the movement vector, where the block vector information is determined by signaling and / or template matching within a predefined search range and / or any predefined implicit or explicit rules. [000178] In another sub-modality of history-based model information, a history-based table (the FIFO table) is created and stores the model information of the previously encoded blocks. The table can be redefined as the beginning and / or end of a CTU, CTU line, slice, image, block, and / or sequence. One or more history-based candidates can be added to the candidate list in order from head to end of the table or from end to head of the table. [000179] In another submodality of paired average model information, the model information for this candidate is derived based on the model information of more than one of the previous candidates in the list. For example, it could calculate the average and / or modify the model parameters of more than one candidate as the model parameters to be applied. For Petition 870250094644, dated 10 / 16 / 2025, page 67 / 111 60 / 94 is another example; you could combine more than one prediction as the final prediction, where each of the multiple predictions is generated by applying one of the models to the list of candidates. [000180] In another sub-mode, the standard template information is added if the list is not complete after all predefined candidates have been entered. Some examples of the standard CCLM template information are shown below. For example, the standard alpha (or denoted α, a, or scaling parameters) are {0, 1 / 8, -1 / 8, 2 / 8, -2 / 8, 3 / 8, -3 / 8, ...}, and the beta (or denoted β, b, or deviation parameter) is based on the selected standard alpha, the average value of the neighboring reconstructed luma sample, and / or the average value of the neighboring reconstructed chroma sample (Cb / Cr). [000181] In another sub-mode, when selecting a candidate from the list and using the candidate's model information for the current block, or when inheriting model information from a previously encoded block (when placing a candidate in the list), only a subset of the model information is inherited. For example, only alpha is inherited. Beta is obtained for the current block through the inherited alpha, the average value of the neighboring reconstructed luma sample, and / or the average neighboring value of the reconstructed chroma sample (Cb / Cr). For example, when inheriting information from the MMLM model, the scale parameters and / or the classification limit are inherited. The deviation parameter in each class is derived according to the inherited classification limit and / or the average value of the neighboring reconstructed luma sample and / or the average value of the neighboring reconstructed chroma sample (Cb / Cr) in each class.If no reconstructed neighboring samples are available in a class, the deviation parameter is inherited directly from the candidate. For example, when inheriting information from the CCCM model, all convolution parameters, deviations, and / or the classification limit are inherited. For example, when inheriting information from the GLM model, if the GLM candidate is the GLM mode of... Petition 870250094644, dated 10 / 16 / 2025, pp. 68 / 111 61 / 94 parameters, all gradient pattern indices and model parameters are inherited; otherwise, if the GLM candidate is the 2-parameter GLM mode, the deviation parameter is derived using the inherited scale parameter, the average value of the reconstructed neighboring luma sample, and / or the average value of the reconstructed neighboring chroma (Cb / Cr) sample. For example, when inheriting chroma fusion model information, the derived MMLM parameters are inherited and used as when inheriting an MMLM candidate for the current block. [000182] In another sub-mode, when selecting a candidate from the list and using the candidate's model information for the current block, all model information is inherited. For example, both alpha and beta are inherited. [000183] In another embodiment, for the current block to be a larger block with width, height, or area greater than a predefined limit, the current block is split into multiple subblocks. For example, the splitting rule is that a minimum block size is predefined, and the current block is split until the width or height of the subblock reaches the minimum block size. For another example, the splitting rule follows the square tree split (4 subblocks) or the binary tree split (2 subblocks). In a sub-emphasis, each subblock derives its own set of candidates or uses its own reference data to obtain the model to generate the prediction. For example, each subblock will have its own list. For another example, the model is derived using current prediction samples, such as offset prediction, model samples, or both. In another embodiment, an implicit rule is defined to select the model (from the list or set) for each subblock.For example, the implicit rule is to use spatial model information for sub-blocks near the upper or left boundary of the current block and / or use temporal model information for the sub-block (e.g., the sub-block in the lower right of the block). Petition 870250094644, dated 10 / 16 / 2025, pp. 69 / 111 62 / 94 current) distant from the upper or left limit of the current block. If the current block is divided into 4 sub-blocks by quadtree, the sub-block in the upper left corner uses the spatial candidate model B2, the sub-block in the upper right corner uses the spatial candidate model B1 or B0, the sub-block in the lower left corner uses the spatial candidate model A1 or A0, and / or the sub-block in the lower right corner uses the temporal candidate model. Sub-blocks without any significant luma and / or cbf residue are ignored. [000184] In another embodiment, when constructing the model list, one or more self-derived cross-component candidates are included, which may correspond to one or more models derived using the methods proposed in Section I and / or Section II.2. The model parameters can be viewed as the weighting configuration in the regression technique. In a sub-embodiment, self-derived cross-component candidates are added only when the set or list contains insufficient inherited candidates or when no inherited candidates are found. The self-derived cross-component candidate refers to one or more models, and the models are used to generate the current block cross-component prediction as follows. The current block cross-component prediction (containing target predicted samples) is formed by combining one or more proposed source terms and the models (referring to a proposed weighting configuration). II.2 Derivation of Cross-Componist Prediction Using Regression Technique [000185] The regression-based technique can be applied to the case of between-component prediction. Expression (6) described earlier can be applied to this case by appropriately defining the contents of sourceTermSet0 (i, j), the contents of sourceTermSet1(i,j) and biasTermSet, as disclosed in Section I.1, Section I.2 and Section I.3, as well as the derivation of the Predictor. Petition 870250094644, dated 10 / 16 / 2025, pp. 70 / 111 63 / 94 for sample (i, j) in Section I.4. [000186] The reference region (i.e., the model) shown in Figure 14 can be applied to the current case. In another embodiment, the current block reference region is the colocalized region of the current block vector (chroma), and the corresponding luminance block reference region is the colocalized region of the corresponding luminance block vector. For example, when deriving the model using the reference data, the corresponding luminance component reference data is the reconstruction of the corresponding luminance component for the reference region, which is the source term, and the chrominance component reference data is the reconstruction of the chrominance component for the reference region to be estimated. The vector colocalization region is a region indicated using a vector in a predefined image, which can be a reference image, the current image, or a colocalized image.In one embodiment, for an inter-coding unit containing luma and chroma blocks, the vector colocalized region of the current block refers to or provides the motion-compensated results for the current chroma block using the motion information (motion vectors and / or reference images) of the current block, and the vector colocalized region of the corresponding luma block refers to or provides the motion of the corresponding luminance component for the reference region. The reference region is the source term, and the chrominance component reference data is the reconstruction of the chrominance component for the reference region to be estimated. The vector colocalization region is a region indicated using a vector in a predefined image, which can be a reference image, the current image, or a colocalized image.In one embodiment, for an inter-coding unit containing luma and chroma blocks, the vector-coalized region of the current block refers to, or provides, the motion-compensated results for the block. Petition 870250094644, dated 10 / 16 / 2025, pp. 71 / 111 64 / 94 of the current chroma using motion information (motion vectors and / or reference images) from the current block, and the corresponding vector-coalized luma block region refers to or provides the motion. The corresponding luminance component for the reference region is the source term, and the chrominance component reference data is the reconstruction of the chrominance component for the reference region, which is the target of the estimate. The vector-coalized region is a region indicated using a vector in a predefined image, which can be a reference image, the current image, or a coalized image.In one embodiment, for an inter-coding unit containing luminance and chrominance blocks, the vector-coalized region of the current block refers to or provides motion-compensated results for the current chroma block using the motion information (motion vectors and / or reference images) of the current block, and the vector-coalized region of the corresponding luma block refers to or provides motion-compensated results for the corresponding luma block using the motion information (motion vectors and / or reference images) of the corresponding luma block.In one embodiment, for IBC or intraTMP (or referred to as TMP), the colocalized region of the current block refers to or provides the motion-compensated results (specifically, block-compensated results) for the current chrominance block, using the motion information (specifically, the block information as block vectors and / or current image) of the current block, and the colocalized region of the corresponding luma block vector refers to or provides the motion-compensated results (specifically, block-compensated results) for the corresponding luma block, using the motion information (specifically, the block information as block vectors and / or current image) of the corresponding luma block. Petition 870250094644, dated 10 / 16 / 2025, pp. 72 / 111 65 / 94 [000187] In another embodiment, the two types of reference region proposed above (e.g., the reference region includes one or more models and a predefined region indicated using a vector) of the current block can be used together. For example, generally, samples in the colocalized region of the current block vector are used as input samples when deriving the model parameters; however, for a smaller block, samples in the neighboring spatial reference region (e.g., a model) are used as additional input samples when deriving the model parameters. [000188] The additional expression for this case is shown below. II.2.1 predc(i,j) = aO· LO + a· LI + a2· L2 + a3· L3 + a4· L4 + a5· L5 + a6· P(i,j)+ a7· bias [000189] In this expression of the target sample being chroma, sourceTermSet0 includes six derivations as L0 to L5 and one derivation P as a non-linear term, sourceTermSet1 is not used and biasTerm refers to another derivation as midValue. L0 to L5 refer to the reconstructed non-subsampled corresponding luma samples referred to by the chrominance to be predicted (i, j) (denoted as the circle in Figure 15). P is generated by any one or more reconstructed non-subsampled corresponding luma samples. For example, (average of the two predefined corresponding luminance samples +1) >> 1) is used and / or P is obtained by following the non-linear term in the CCCM method, for example, the CCCM method in the intra process. The two predefined corresponding luminance samples refer to the two upper and lower samples next to the circle in Figure 15. [000190] The model parameters a0 to a7 are derived using a regression method and / or without using division operations. Before deriving the parameters, the proposed deviations are used to fit the input samples. [000191] In another modality, the following is also used Petition 870250094644, dated 10 / 16 / 2025, pp. 73 / 111 66 / 94 sourceTermSet1. For example, one or more additional derivations for sourceTermSet1 refer to the predicted initial sample (i, j) for the current block and / or a pattern around (i, j) generated using the prediction mode for the current block. For the inter encoding unit containing luma and chroma blocks, the predicted initial sample (i, j) refers to the motion-compensated results using the motion information (motion vectors and / or reference images) of the current block. For IBC or intraTMP, the predicted initial sample (i, j) refers to the motion-compensated results (specifically, block-compensated results) using the motion information (specifically, block information such as block vectors and / or current image) of the current block. The additional derivations are derived using the spatially neighboring reference region of the current block. [000192] In another embodiment, sourceTermSet0 or sourceTermSet1 may include gradient terms in other examples. In another embodiment, the derivation of the prediction between components may follow unified methods in the intra process. In another embodiment, more variations of the expression may reference section I. For example, the number of derivations and / or the source terms of the expression may be different or unified with any predefined cross-component intra mode. II.3 MODEL INFORMATION SIGNALING [000193] When the methods proposed in section II.5 (e.g., inter CCLM) or inter CCCM do not apply, the prediction of the current block is made from the original prediction (e.g., inter original prediction). [000194] In another modality, the choice of whether or not to apply the methods proposed in section II.5 depends on the signaling. For example, whether or not to apply inter CCLM depends on the signaling. For another example, whether or not to apply inter Petition 870250094644, dated 10 / 16 / 2025, pp. 74 / 111 67 / 94 CCCM depends on signage. [000195] In a submode, signaling refers to a coded indicator of TU, ​​TB, CU, and / or CB level. The flag may or may not depend on the context to be encoded. Taking the TU / TB indicator as an example, the indicator is signaled only if the luma Cbf of TU / TB is non-zero (e.g., coded block indicator for luma TU / TB being equal to true and / or the activation indicator for inter mode being true). Taking the CU / CB indicator as an example, the indicator is signaled only if the luma Cbf of CU / CB is non-zero, e.g., the coded block indicator for luma CU / CB being equal to true and / or the activation indicator for inter mode being true. The activation flag for inter mode means that the predMode of CU is MODE_INTER when the proposed inter CCLM (or inter CCCM) is supported for all inter modes. When the proposed inter CCLM (or inter CCCM) is supported for IBC.The activation indicator for IBC is checked first, and the signaling for CCLM inter (or CCCM inter) is encoded or decoded in response to the CU's predMode being MODE_IBC. [000196] In another submode, an indicator is signaled in the bitstream to indicate whether inter-CCLM (or inter-CCCM) should be applied or not. For example, the indicator is context-encoded. In another example, only one context is used when encoding the indicator. In yet another example, multiple contexts are used to encode the indicator and / or the selection of contexts depends on block width, block height, block area, or neighbor mode information. [000197] In another sub-mode, when the signaling indicates applying the methods proposed in section II.5 (e.g., inter CCLM) or inter CCCM, additional signaling is used to determine how to obtain the cross-component model for the current block. In one case, select one or more models from the total number of candidates (e.g., CCLM_LT, CCLM_L, CCLM_T, MMLM_L, Petition 870250094644, dated 10 / 16 / 2025, pp. 75 / 111 68 / 94 MMLM_T, MMLM_L, or any subset or extension of the modes mentioned above). For example, if an LM mode is selected, the LM prediction is generated by the selected LM. For another example, if more than one LM mode is selected, the LM prediction is generated by combining prediction hypotheses from multiple LM modes. For yet another example, additional signaling refers to an index in the bitstream, which can be truncated unary encoding with and / or without contexts. [000198] In another sub-mode, when the signaling indicates applying the methods proposed in section II.5 (e.g., inter CCLM) or inter CCCM, one or more models from the total number of candidates (e.g., CCLM_LT, CCLM_L, CCLM_T, MMLM_L, MMLM_T, MMLM_L or any subset or extension of the modes mentioned above) is / are implicitly selected (or predefined) to be used in the generation of inter-component prediction (e.g., prediction for inter CCLM or inter CCCM). For example, CCLM_LT is used to generate LM predictions for inter-CCLM. For example, MMLM_LT is used to generate LM predictions for inter-CCLM. Similar rules can be applied to inter-CCCM, alternatively, using convolutional models. - For another example of inter CCLM, the predefined rule depends on the block width, block height, or block area. Similar rules can be applied to inter CCCM using, as an alternative, convolutional models. - The boundary matching setting (used as the default rule) can only be applied when the width, height, or area of ​​the block is greater than a limit. - The boundary matching setting (used as the default rule) can only be applied when the width, height, or area of ​​the block is less than a certain limit. Petition 870250094644, dated 10 / 16 / 2025, pp. 76 / 111 69 / 94 - When the width, height, or area of ​​the block is less than a threshold, the selected LM mode(s) is / are inferred as any one (or more than one) LM mode(s) from the total number of candidate LM modes. The selected LM mode is set to CCLM_LT. The selected LM mode is set to MMLM_LT. - For another example, the predefined rule depends on the boundary matching configuration. (Details of the boundary matching configuration can be found in the section titled “Boundary Matching Configuration.” The candidate mode used in the Boundary Matching Configuration section refers to each candidate LM mode for inter-CCLM. The prediction of a candidate mode used in the Boundary Matching Configuration section refers to the prediction generated by each candidate LM mode or refers to the combined prediction of each candidate LM mode and the original inter-CCLM). Similar rules can be applied to inter-CCCM using convolutional models as an alternative. [000199] In another embodiment, the methods proposed in section II.5, for example, inter CCLM (or inter CCCM), can only be supported when the current block size conditions are met. [000200] In a submode, the size condition is that the block width, block height, or block area is greater than a predefined limit. The predefined limit can be a positive integer, such as 8, 16, 32, 64, 128, 256, etc. [000201] In another sub-mode, the size condition is that the block width, block height, or block area is less than a predefined limit. The predefined limit can be a positive integer, such as 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, etc. [000202] In another mode, the inter mode used in the block Petition 870250094644, dated 10 / 16 / 2025, pp. 77 / 111 70 / 94 inter-interface depends on an activation indicator. For example, if the inter-interface mode is regular merge, the activation indicator is referred to as the regular merge indicator. For another example, if the inter-interface mode is CIIP, the activation indicator is referred to as the CIIP indicator. For another example, if the inter-interface mode is CIIP PDPC, the activation indicator is referred to as the CIIP PDPC indicator. For another example, the activation indicator indicated as enabled (equal to 1) means that the corresponding inter-interface mode is applied to the current block. For another example, the activation indicator indicated as disabled (equal to 0) means that the corresponding inter-interface mode is not applied to the current block. For another example, the activation indicator is signaled in the bitstream and / or inferred in some cases. On the other hand, the signaling of the activation indicator depends on the width, height, or area of ​​the block. [000203] In another mode, the prediction from the inter can be adjusted by neighboring reconstructed samples and by a predefined weighting scheme. For example, when the current block is merged, the merge prediction is combined with neighboring reconstructed samples. In another example, the proposed scheme is activated depending on the CIIP PDPC indicator. (The CIIP PDPC indicator can be signaled when the CIIP indicator is indicated as activated). For another example, the predefined weighting scheme follows the PDPC weighting. - The interpredictor of the regular blending mode is refined using the reconstructed samples above Rx-i and left Ri,y; The derivation of nScale and wT & wL is the same as in intraplanar mode. - wT = 32 >> ( ( y<<1 ) >> nScale) - wL = 32 >> ( ( x<<1 ) >> nScale) - nScale = (floorLog2(width) + floorLog2(height) - 2) >> 2; - CIIP PDPC: Petition 870250094644, dated 10 / 16 / 2025, pp. 78 / 111 71 / 94 If LMCS is enabled, the interpredictor is calculated in the mapped domain. Pred(x,y) = ((((W x R^ + wLx R_1,y+ 32) » ó) « ó) + (64 — wT — wL) x Fwd(predlnter(x,y)) + 32) » 6; Otherwise, the interpredictor is calculated in the original domain. Pred(x,y) = ((((wT x Rx,_1+ wLx R-1y+ 32) » 6) « 6) + (64 — wT — wL) x predlnter(x, y) + 32) » 6. - When the CIIP flag is true, the CIIP flag PDPC is flagged to indicate whether CIIP PDPC should be used. [000204] In another embodiment, the inter-original prediction (generated by motion compensation) is used for luma, and the chroma component predictions are generated by CCLM and / or any other LM modes and / or any other cross-component modes or models, for example, the modes or models used in Section I, Section II.2 and / or Note A. [000205] In a submodality, the current UC is viewed as an inter, intra, or a new type of prediction mode (neither intra nor inter). [000206] In another embodiment, a derivation of the LM or cross-component mode is disclosed as described below (referred to as Note A): One or more LM modes (or cross-component modes), which may or will be used to generate one or more prediction hypotheses for the LM-assisted Angular or Planar Mode and / or methods proposed in section II.5 (e.g., interCCLM) and / or MH CCLM, are selected from a predefined set or list, for example, a set or list of merge candidates (where the set or list is called modelList). A modelIdx is signaled to select a candidate from the set or list of candidates (modelList), and the selected candidate is used for the current block. The modelList contains one or more candidates, where each candidate refers to a piece of model information (or mode of Petition 870250094644, dated 10 / 16 / 2025, pp. 79 / 111 72 / 94 cross-component). If there is only one candidate in the set or list (the size of the set or list is only 1), the modelIdx is not flagged and / or can be inferred as 0 or a default value, which means that it must be selected implicitly. [000207] In one embodiment, when constructing modelList, one or more predefined candidates are added. The predefined candidates may include any subset / extension of the following candidates. In one case, all candidates in modelList are self-derived. For example, any or all subsets of the CCLM family, MMLM family, CCCM family, any cross-component modes in Section I, self-derived candidates in Section II.1 and / or Section II.2. In another case, some candidates in modelList are self-derived and some candidates in modelList are inherited, for example, inherited candidates in Section II.1. - CCLM family: CCLM_LT, CCLM_L, CCLM_T. - MMLM family: MMLM_LT, MMLM_L, MMLM_T. - CCCM family: CCCM_LT, CCCM_L, CCCM_T. [000208] Any cross-component modes mentioned in Section I, Section II.1 and / or Section II.2. [000209] The methods proposed above can also be applied to IBC blocks or blocks with any IBC submodes (e.g., IBC merge or IBC AMVP or any IBC mode under the IBC syntax). (inter in this invention can be changed to IBC). That is, for chroma components, block vector prediction can be combined with or replaced by cross-component prediction. II.4 Using Model Information to Generate Prediction Hypotheses [000210] In one embodiment, a prediction- or reconstruction-based model is used to generate a predictive hypothesis for the current chroma component. Petition 870250094644, dated 10 / 16 / 2025, pp. 80-111 73 / 94 [000211] In one embodiment of a linear prediction-based model, the derived model parameters are applied to the predicted samples for the first component (Y) to obtain the predicted samples for the second or third component. Let us take as an example the parameters of the LM model, including a and b. P(i,j) = a · pred'L(i,j) + b. [000212] The predicted samples for the first component are subsampled using subsampling filters (which can be fixed to a predefined filter or selected from several candidate filters). For example, the subsampling filters follow the original LM design. In another example, the subsampling filters will not access neighboring predicted or reconstructed samples. At the current block boundary, if neighboring samples need to be the input samples for the subsampling filters, populated predicted values ​​will be used instead, starting from the current block boundary. [000213] In another submodality of a reconstruction-based linear model, the parameters of the derived model are applied to the reconstructed samples for the first component (Y) to obtain the predicted samples for the second or third component. Let us take as an example the parameters of the LM model, including a and b. P(i,j) = a · reco'L(i,j) + b. [000214] The reconstructed samples for the first component are subjected to subsampling with subsampling filters (which can be fixed to a predefined filter or selected from several candidate filters). For example, the subsampling filters follow the original LM design. In another example, the subsampling filters will not access predicted or reconstructed neighboring samples. At the current block boundary, if neighboring samples need to be the input samples for the subsampling filters, predicted values ​​populated from the block boundary will be used instead. Petition 870250094644, dated 10 / 16 / 2025, pp. 81 / 111 74 / 94 current. [000215] The prediction- or reconstruction-based convolution model is similar to the methods proposed for the prediction- or reconstruction-based linear model. The main difference is that the model coefficient pattern follows CCCM (not CCLM), and luminance samples may or may not be subsampled first. If subsampling is not applied to the luminance samples, more derivations (model coefficients) can be used to access the non-subsampled luminance samples. [000216] In another embodiment, multiple cross-component prediction hypotheses (MH) are combined or multiple models are used to generate a prediction hypothesis for the current block. Each cross-component method, for example, a CCLM method, is suitable for different scenarios. For some complex features, combined prediction may result in better performance. Therefore, cross-component predictions of multiple hypotheses, for example, including CCLM prediction (a single-hypothesis CCLM example to predict Cb or Cr samples based on Y samples), are proposed to combine the predictions of several cross-component methods, for example, CCLM methods. The CCLM methods to be combined may be (but are not limited to) the CCLM methods mentioned above and / or the methods disclosed in Section I, Section II.1, Section II.2 and / or Note A. A weighting scheme is used for the combination. [000217] In one embodiment, the weightings for different CCLM methods are predefined in the encoder and decoder. [000218] In another modality, the weightings vary based on the distance between the sample (or region) positions and the reference sample positions. [000219] In another mode, the weightings depend on neighboring encoding information. [000220] In another modality, a weighting index is Petition 870250094644, dated 10 / 16 / 2025, pp. 82 / 111 75 / 94 flagged or analyzed. Codewords can be fixed or vary adaptively. For example, codewords vary with model-driven methods. [000221] Similar rules can be applied to CCCM using convolutional models instead, or MH methods can be applied between CCLM and CCCM. For example, one hypothesis is from CCLM and another hypothesis is from CCCM. [000222] In another sub-modality, one or more assumptions of predictions between components can be combined with one or more assumptions of prediction modes between to form the final prediction of the current block. [000223] Next, a CCLM flow based on prediction is shown. - Improve chrominance prediction by linearly predicting chrominance samples from luminance samples; The linear prediction method can be one of the following: CCLM_LT, CCLM_L, CCLM_T - MMLM_LT, MMLM_L, MMLM_T - Steps: - Step 1: Derive the linear model from reconstructed samples of neighboring luma and chroma; Step 2: Apply the derived linear model to the predicted actual luma samples to obtain the predicted actual chroma samples: - predccL»(tj) = α · predL'(i,j) + β; - predL'(i,j): predicted samples of actual luma with subsampling; - Padding is used within the current block's boundary. II.5 CCLM FOR INTERBLOCK [000224] The CCLM for interblock can also be called interCCLM and CCLM can be extended to any LM mode (or any cross-component mode) or replaced by any LM mode (or any cross-component mode, for example, Petition 870250094644, dated 10 / 16 / 2025, pp. 83 / 111 76 / 94 CCCM, or any model or weighting proposed in Section I and / or Section II.1 and / or Section II.2 and / or Section II.3, for example, Note A). [000225] In the overview section, CCLM is used for intrablocks to improve intra-chroma prediction. For an interblock, chroma prediction may not be as accurate as luma prediction. Possible reasons are listed below: The motion vectors for chroma components are inherited from the luma (for example, the chroma does not have its own motion vectors). Few coding tools are designed to improve inter-chroma prediction. [000226] Therefore, an alternative way of applying CCLM to interblocks is proposed. With this proposed method, chroma prediction for the interblock can be improved according to luminance. [000227] In one embodiment, for chroma components, in addition to the original inter prediction (generated by motion compensation which may be uni-prediction and / or bi-prediction, or which may be only uni-prediction), one or more prediction hypotheses (generated by CCLM and / or any other LM modes and / or any cross-component mode, such as CCCM, or any model or weighting proposed in Section I and / or Section II.1 and / or Section II.2 and / or Section II.3, such as Note A) are used to produce the actual prediction. [000228] In a submodality, the current prediction is the weighted sum of the inter-prediction and the CCLM prediction. The weightings are designed according to neighboring coding information, sample position, block width, height, or area. For example, for a small block (e.g., area < limit), the weightings for CCLM prediction are higher than the weightings for inter-prediction. - For another example, when most of the encoded blocks Petition 870250094644, dated 10 / 16 / 2025, pp. 84 / 111 77 / 94 neighbors are intra-blocks or a cross-component coded block, for example, CCLM coded blocks, the weights for CCLM prediction are higher than the weights for inter-prediction. On the other hand, when most of the neighboring coded blocks are inter-blocks, the weightings for inter-prediction are higher than the weightings for CCLM prediction. On the other hand, the weightings are fixed values ​​for the entire block. [000229] In another embodiment, the inter prediction can be generated by any inter mode mentioned in the introduction or documents referenced above. For example, the inter mode can be the regular fusion mode. Alternatively, the inter mode can be the CIIP mode. Alternatively, the inter mode can be CIIP PDPC. Alternatively, the inter mode can be GPM or any GPM variations (e.g., intra GPM, which forms the final prediction by combining the intra prediction and the inter prediction with a weighting according to the geometric partition). [000230] In one submodality, the regular merge mode is a merge candidate selected from the list of merge candidates with a signaled merge index. In another submodality, the regular merge mode may be MMVD. [000231] In another sub-modality, the LM mode used in CCLM inter is prediction-based LM. [000232] In another embodiment, inter CCLM is supported only when any (or more than one) of the predefined inter modes is used for the current block, or inter CCLM is supported when any (or more than one) of the predefined inter mode activation indicators is indicated as activated. The significance of supporting inter CCLM is that the prediction of the current block can be chosen between applying inter CCLM or not applying inter CCLM. CCLM can be extended to any LM mode (or any cross-component mode) or replaced by any Petition 870250094644, dated 10 / 16 / 2025, pp. 85 / 111 78 / 94 LM mode (or any cross-component mode, for example, CCCM, or any model or weighting proposed in Section I and / or Section II.1 and / or Section II.2 and / or Section II.3, for example, Note A). [000233] When applying inter CCLM, the current block prediction is generated by: - In a submodality: mixing one or more prediction hypotheses (generated by CCLM and / or any other LM modes and / or any cross-component mode) with the original inter-prediction: - Combining chroma prediction for the inter-existing mode and LM prediction. - Mixture: Predfinal = ( wInter * PredInter + wLM * PredLM + 2 ) >> 2. - Weighting rule: wInter and wLM, for example. - If both the top and left parts are intra (or any cross-component mode), (wInter, wLM) = (1, 3) Otherwise, if one of the top left parts is intra, (wInter, wLM) = (2, 2). Otherwise, (wInter, wLM) = (3, 1). - For another example, the weighting follows the CIIP weighting rules. For example, predInter = inter prediction after OBMC (if OBMC is used). - For another example, predInter = inter prediction before OBMC (OBMC can be applied after mixing). [000234] In another submodality, the original inter-prediction is replaced by one or more prediction hypotheses (generated by CCLM and / or any other cross-component modes). [000235] In another sub-mode, the CCLM mode can be inherited or modified from neighboring blocks or any encoded blocks. For example, luma prediction comes from inter-encoding tools, the prediction of Petition 870250094644, dated 10 / 16 / 2025, page 86 / 111 79 / 94 chroma is derived using luma prediction or reconstruction with the CCLM model, and the CCLM model is inherited or modified from neighboring blocks. To inherit or modify CCLM models from neighboring blocks or any encoded blocks, a candidate list or historical list is created to include CCLM models used in adjacent and non-adjacent neighboring blocks or positions. It may also include CCLM models used in previous encoded images or slices. Then, an index is used to indicate which model from the list is inherited or modified to generate the current chroma prediction. [000236] For another example, if the CCLM mode is used to generate the chroma prediction samples and the luma prediction is from an inter-coding tool, an indicator is used to indicate whether the CCLM model used for chroma prediction is inherited from the CCLM models used in previous coded blocks or the CCLM model is from a predetermined CCLM mode. If the CCLM model is inherited from the CCLM models used in previous coded blocks, an index is used to indicate which model in the list is inherited or modified. Otherwise, a predetermined CCLM mode is used to implicitly derive the CCLM model for the current chroma prediction. CCLM can be extended to any LM mode (or any cross-component mode) or replaced by any LM mode (or any cross-component mode, e.g., CCCM, or any model or weighting proposed in Section I and / or Section II.1 and / or Section II.2 and / or Section II.3, such as Note A).For example, the cross-component model is selected between using a legacy model or using a derived model automatically. III Angular or Planar Mode Assisted by LM [000237] For traditional intra-prediction modes (e.g., angular, DC, and planar intra-prediction modes), the reference samples are derived from reconstructed upper and left neighboring samples. Therefore, the Petition 870250094644, dated 10 / 16 / 2025, page 87 / 111 80 / 94 intra-prediction accuracy decreases for lower right samples within the current block. In this section, LM or any cross-component mode is used to improve prediction over traditional intra-prediction modes. [000238] In one embodiment, the prediction for the current block is formed by a weighted sum of one or more prediction hypotheses from traditional intra-prediction modes and one or more prediction hypotheses from LM modes (or cross-component modes). - In a sub-modality, equal weightings are applied to both. - In another sub-mode, the weightings vary with neighboring coding information, sample position, block width, height, mode, or area. For example, when the sample position is far from the upper left region, the weighting for prediction using traditional intra-prediction methods decays. Some examples are presented below. One possible rule related to the sample position is described below. - When the sample position is further away from the reference samples, the weighting for prediction by other intra-prediction modes, for example, traditional intra-prediction modes, decreases. Another possible rule related to neighboring encoding information is described below. - When more neighboring blocks (left, above, left above, right above, and / or left below) are encoded with a specific mode (e.g., Mode A), the weighting for Mode A prediction increases. For example, Mode A refers to a specific cross-component mode, such as CCCM_LT, or a specific cross-component family, such as the CCCM family (including CCCM_LT, CCCM_L, and / or...). Petition 870250094644, dated 10 / 16 / 2025, pp. 88 / 111 81 / 94 CCCM_T) and / or CCLM family (including CCLM_LT, CCLM_L and / or CCLM_T) and / or MMLM family (including MMLM_LT, MMLM_L and / or MMLM_T). A weighting set is predefined to include several weighting candidates, such as {1, 3}, {3, 1} and / or equal weighting {2, 2} for one prediction from a traditional intra-component prediction mode and the other prediction from a cross-component mode. When all or most neighboring blocks are Mode A coded, the weighting candidate with the highest weighting for the cross-component prediction mode is used. When only some or one of the neighboring blocks are Mode A coded, the weighting candidate with an equal weighting is used. When few or none of the neighboring blocks are Mode A coded, the weighting candidate with the lowest weighting for the cross-component prediction mode is used.For example, neighboring blocks include any subset of the coded blocks that are spatially adjacent to the upper or left boundary of the current block. In this case, neighboring blocks can refer to the upper neighboring block (located at the top right corner of the current block) and the left neighboring block (located to the left of the lower left corner of the current block). Alternatively, neighboring blocks include any subset of the coded blocks located within a predefined spatial range close to the upper or left boundary of the current block. In this case, neighboring blocks can be adjacent or non-adjacent to the current block. Another possible rule related to the sample position is described below. The current block is divided into several regions. Sample positions within the same region share the same weighting. If the current region is close to the reference L neighbor, the weighting for predicting other intra-prediction modes, for example, traditional intra-prediction modes, is greater than Petition 870250094644, dated 10 / 16 / 2025, pp. 89 / 111 82 / 94 which is the weighting for predicting cross-component modes, for example, CCLM. The following shows some possible ways to split the current block. - (width to height ratio close to or exactly 1:1): the distance between the current region and the reference neighbor L to the left and above is considered. - (width > n * height, where n can be any positive integer): the distance between the current region and the reference neighbor L above is considered. - (height > n * width, where n can be any positive integer): the distance between the current region and the left reference neighbor L is considered. [000239] Further embodiments of the invention are presented below. In another subemphasis, the mode to propagate to the current block (using chroma fusion or termed Angular or Planar LM-assisted mode, methods proposed in section II.5 (e.g., inter CCLM) or MH CCLM) is defined or stored. For example, defining or storing it as the inherited mode, such as CCLM, MMLM, CCCM, or GLM. [000240] In another submode, the candidate types are aligned with the candidate types for the merge mode. For example, the candidate types in modelList are aligned with the candidate types for the merge mode. [000241] In another submodality, when constructing the set or list, pruning operations are applied to avoid duplicate candidates in the list. [000242] In another submodality, the modelIdx signaling depends on the context encoding, block width, block height, block area, and / or explicit syntax, such as SPS, PPS, slice, CTU, image, sequence, and / or block-level signaling. [000243] In another sub-modality, the selection of candidates (from the set or list, for example, modelList) for the current block depends on a predefined process. For example, the process Petition 870250094644, dated 10 / 16 / 2025, pp. 90 / 111 83 / 94 predefined is a model-based mode derivation method, for example, a TIMD or DIMD type method. For example, the predefined process depends on the model neighbor of the current block. For example, the predefined process depends on the mode information of neighboring blocks. If most of the predefined neighboring blocks use a specific mode, the first candidate (in the list) that refers to the specific mode is selected. The modelIdx signaling is ignored (it is not needed for signaling). The methods proposed here can be applied to Note A. [000244] In one embodiment, chroma refers to the Cb and / or Cr component(s). In another sub-emphasis, only one of the Cb and Cr pieces of information is used. In yet another sub-emphasis, the chroma information is from both Cb and Cr. For example, the reconstructed neighboring Cb and Cr samples are weighted and then used as inputs to derive the model parameters. In another example, the reconstructed Cb and Cr samples in the chroma block (placed alongside the current luminance block) are weighted and then used to derive the predictors in the current luminance block. [000245] In another embodiment, for the current luminance block, the prediction (generated by the proposed inverse LM) can be combined with one or more prediction hypotheses (generated by one or more other intra-prediction modes, e.g., traditional intra-prediction modes). The proposed inverse LM is described below. For the CCLM mode, as previously disclosed in the background section, reconstructed luma samples are used to derive the predictors in the chroma block. In this disclosure, the inverse LM is proposed to use chroma information to derive the predictors in the luma block. When supporting the inverse LM, the chroma components are encoded or decoded (signaled or parsed) before the luma component. In one embodiment, the chroma information refers to the chroma samples. Petition 870250094644, dated 10 / 16 / 2025, pp. 91 / 111 84 / 94 reconstructed. When deriving model parameters for inverse LM, reconstructed neighboring chroma samples are used as X, which is used to predict luma when deriving the linear model, and reconstructed neighboring luma samples are used as Y, which is used as the target when deriving the linear model to predict luma. Furthermore, the reconstructed samples in the chroma block (placed in the current luma block) and the derived parameters are used to generate the predictors in the current luma block. An alternative approach is that the information in this mode can refer to predicted samples. [000246] In a submodality, other intra prediction modes may refer to angular, DC, planar, MIP, ISP, MRL intra prediction modes, any other existing intra modes (supported in HEVC / VVC) and / or any other intra prediction modes. [000247] In another sub-modality, when combining multiple prediction hypotheses, the weighting for each hypothesis can be fixed or adaptively altered. For example, equal weightings are applied to each hypothesis. In another example, the weightings vary with neighboring coding information, sample position, block width, height, prediction mode, or area. Some examples of using neighboring coding information are presented below: One possible rule related to the sample position is described below. - When the sample position is further away from the reference samples, the weighting for prediction by other intra-prediction modes, for example, traditional intra-prediction modes, decreases. Another possible rule related to neighboring encoding information is described below. - When more neighboring blocks (left, above, left above, right above, and / or left below) are encoded with Petition 870250094644, dated 10 / 16 / 2025, pp. 92-111 85 / 94 a specific mode (for example, Mode A), the weighting for Mode A prediction becomes higher. Another possible rule related to the sample position is described below. The current block is divided into several regions. Sample positions within the same region share the same weighting. If the current region is close to the reference L neighbor, the weighting for predicting other intra-prediction modes, for example, traditional intra-prediction modes, is greater than the weighting for predicting CCLM. Some possible ways to divide the current block are shown below. - (width to height ratio close to or exactly 1:1): the distance between the current region and the reference neighbor L to the left and above is considered. - (width > n * height, where n can be any positive integer): This is considered the distance between the current region and the upper reference neighbor L. - (height > n * width, where n can be any positive integer): This is considered the distance between the current region and the left reference neighbor L. IV. BOUNDARY MATCHING CONFIGURATION [000248] When the boundary matching configuration is used, a boundary matching cost for a candidate mode refers to the discontinuity measurement (including upper boundary matching and / or left boundary matching) between the current prediction (i.e., the predicted samples within the current block) generated from the candidate mode and the neighboring reconstruction (i.e., the reconstructed samples within one or more neighboring blocks), as shown in Figure 16, where predi,j refers to a predicted block, recoi,j refers to a neighboring reconstructed block, and block 1610 (as shown in a thick-lined box) matches the current block. Upper boundary matching Petition 870250094644, dated 10 / 16 / 2025, pp. 93 / 111 86 / 94 means the comparison between the current upper predicted samples and the neighboring upper reconstructed samples, and the left limit match means the comparison between the current left predicted samples and the neighboring left reconstructed samples. [000249] In one mode, the candidate mode with the lowest boundary matching cost is applied to the current block. [000250] In another embodiment, the limit matching cost for Cb and Cr can be added together to be the limit matching cost for chroma, so that the candidate mode selected for Cb and Cr will be shared. That is, the candidate mode selected for Cb and Cr will be the same. [000251] In another embodiment, the candidate modes selected for Cb and Cr depend on the boundary matching costs for Cb and Cr, respectively, so the candidate modes selected for Cb and Cr may be the same or different. [000252] In one embodiment, a predefined subset of the current prediction is used to calculate the boundary matching cost. n rows of the upper boundary within the current block and / or m rows of the left boundary within the current block are used. (Additionally, n2 rows of the upper neighbor reconstruction and / or m2 rows of the left neighbor reconstruction are used). [000253] In an example of calculating a boundary matching cost, n = 2, m = 2, n2 = 2 and m2 = 2: cost = ^ block width (|a * predXi0- b * predx^ c * recox,_1| + |d * recox_1e * predxi0— f* recox_2\) + £ block height y=o (|^ * predOy— h* pred1y— i * reco_1yl + \j * reco_1y— k * predOy— l * reco_2 y[) . [000254] In the equation above, the weights (a, b, c, d, e, f, g, h, i, j, k, l) can be any positive integers, such as a = 2, b = 1, c = 1, d = 2, e = 1, f = 1, g = 2, h = 1, i = 1, j = 2, k = 1 and l = 1. Petition 870250094644, dated 10 / 16 / 2025, pp. 94-111 87 / 94 [000255] In another example of calculating a boundary matching cost, nm = 2, n2 and m2 cost = Σ block width (]a*predx,0x=ob * predxl— c * recox_l\) [000256] +Σ height of the block Ç\g*pred0,yy=o — h* predly— i * reco_ly\). In the equation above, the weights (a, b, c, gei) can be any positive integers, such as 2, b 1, g = 2, h 1. [000257] In another example of calculating a boundary matching cost, n 1, n2 m2 2: cost = Σ block width (\d*recox_i x=oe * predxi0- f * recoXi_2|) [000258] block height (\j * reco_ y=oi,yk *pred0l*reco_2,y[). In the equation above, the weights (d, e, f, jel) can be any positive integers, such as d = 2, e 1, j 2, k 1. [000259] In another example of calculating a boundary matching cost, n 1, n2 m2 1: cost = Σ block width x=o + Σ block height lg * predOiy y=o reco_lyl. [000260] In the equation above, the weights (a, c, g and g) i) can be any positive integers, such as 1, c 1 g 1. [000261] In another example of calculating a boundary matching cost, n = 2, m = 1, n2 = 2 and m2 = 1: cost Σblock width x=o (\a*predXt0— b * predxl— c * recox_ll + ld* recox_l — e* predXt0— f* recoil) Σblock height y=o (|g * predOy — i * reco_ly[). [000262] In the equation above, the weights (a, b, c, d, e, f, gei) can be any positive integers, such as a = 2, b Petition 870250094644, dated 10 / 16 / 2025, pp. 95-111 88 / 94 = 1, c = 1, d = 2, e = 1, f = 1, g = 1 and i = 1. [000263] In another example of calculating a boundary matching cost, n = 1, m = 2, n2 = 1 and m2 = 2: cost Σ block width x=o (\a*predx,0 - c*recox,_i|) Σ block height y=o (\^ * predOy — h *predly— i * reco_ly\ + \j * reco_ly— k * predOy— I * reco_2y\). [000264] In the equation above, the weights (a, c, g, h, i, j, kel) can be any positive integers, such as a = 1, c = 1, g = 2, h = 1, i = 1, j = 2, k = 1 and kel = 1. The following examples for n can also be applied to n2 and m2. [000265] For another example, n can be any positive integer, such as 1, 2, 3, 4, etc. [000266] For another example, m can be any positive integer, such as 1, 2, 3, 4, etc. [000267] For another example, ne / or m vary with the width, height, or area of ​​the block. In one embodiment, m becomes larger for a larger block (e.g., area > limit2). For example: - Limit2 = 64, 128 or 256. - When the area > limit2, m is increased to 2. (Originally, m is 1). - When the area > limit2, m is increased to 4. (Originally, m is 1 or 2). [000268] In another example, m becomes larger and / or n becomes smaller for a taller block (for example, height > limit2 * width). For example: Limit 2 = 1, 2, or 4. - When height > limit2 * width, m is increased to 2. (Originally, m is 1). - When height > limit2 * width, m is increased to 4. (Originally, m is 1 or 2). [000269] In another mode, n increases to a larger block Petition 870250094644, dated 10 / 16 / 2025, pp. 96 / 111 89 / 94 (area > limit2). - Limit2 = 64, 128 or 256. - When the area > limit2, n is increased to 2. (Originally, n is 1). - When the area > limit2, n is increased to 4. (Originally, n is 1 or 2). [000270] In another embodiment, n increases and / or m decreases for a wider block (width > limit2 * height). For example, [000271] Limit2 = 1, 2 or 4. [000272] When width > limit2 * height, n is increased to 2. (Originally, n is 1). [000273] When width > limit2 * height, n is increased to 4. (Originally, n is 1 or 2). [000274] The methods proposed in this invention can be activated and / or deactivated according to implicit rules (e.g., block width, height, or area) or according to explicit rules (e.g., syntax at the block, slice, image, SPS, or PPS level). [000275] The term block in this invention may refer to TU / TB, CU / CB, PU / PB or CTU / CTB. [000276] The term LM in this invention can be viewed as a type of CCLM / MMLM mode or any other extension or variation of CCLM (for example, the CCLM extension or variation proposed in this invention). One variation is MMLM, which uses boundaries to decide different models for different samples in the current chromatic component. Another variation is that, for Cb (or Cr), the model parameters are derived from several colocalized luminance blocks. The following shows more possible variations. The variations of CCLM here mean that some optional modes can be selected when the block indication refers to the use of one of the cross-component modes (for example, CCLM_LT, MMLM_LT, CCLM_L, CCLM_T, MMLM_L, MMLM_T and / or an intra prediction mode, which is not one of the traditional DC, planar and angular modes). Petition 870250094644, dated 10 / 16 / 2025, pp. 97-111 90 / 94 for the current block. The following shows an example of the cross-component convolutional mode (CCCM) as an optional mode. When this optional mode is applied to the current block, cross-component information with a model, including the nonlinear term, is used to generate the chroma prediction. The optional mode can follow the CCLM model selection; therefore, the CCCM family includes CCCM_LT, CCCM_L, and / or CCCM_T. [000277] The methods proposed (for CCLM) in this invention can be used for any other cross-component modes. For example, any other cross-component modes or designs used in Section I, Section II.2, Section II.1 and / or Note A. [000278] Any of the cross-component prediction derivation methods using reference data of the corresponding luminance and chrominance components can be implemented in encoders and / or decoders. For example, any of the proposed methods can be implemented in an encoder's inter or intra or IBC candidate derivation module or prediction or transformation or fusion module and / or in a decoder's inter or intra or IBC candidate derivation module or prediction or transformation or fusion module.Alternatively, any of the proposed methods can be implemented as a circuit coupled to the inter or intra or IBC candidate derivation module or prediction or transformation or fusion of the encoder and / or to the inter or intra or IBC candidate derivation module or prediction or transformation or fusion of the decoder, in order to provide the necessary information to the inter or intra or IBC candidate derivation module or prediction or transformation or fusion. For example, any of the proposed methods can be implemented in an inter or intra or prediction module (e.g., Intra Pred. 110 in Figure 1A) of an encoder and / or in an inter or intra or prediction module (e.g., Intra Pred. 150 in Figure 1B) of a decoder. Petition 870250094644, dated 10 / 16 / 2025, pp. 98 / 111 91 / 94 Alternatively, any of the proposed methods can be implemented as a circuit coupled to the inter- or intra- or prediction module of the encoder and / or to the inter- or intra- or prediction module of the decoder, in order to provide the necessary information to the inter- or intra- or prediction module. [000279] Figure 17 illustrates a flowchart of an exemplary video coding system that derives prediction between components using reference data, for example, compensated prediction or model reconstruction, of the corresponding luminance and chrominance components, according to an embodiment of the present invention. The steps shown in the flowchart can be implemented as executable program codes on one or more processors (e.g., one or more CPUs) on the encoder side and / or the decoder side. The steps shown in the flowchart can also be implemented based on hardware, such as one or more electronic devices or processors arranged to execute the flowchart steps.According to this method, the input data associated with a current block in a current image comprising a first color component and a second color component are received in step 1710, where the input data comprises pixel data to be encoded on the encoder side or data associated with the current block to be decoded on the decoder side, and where the current block comprises a first color block and a second color block. The application of a target mode to the current block is determined in step 1720. If the target mode is applied to the current block (i.e., the Yes path of step 1720), steps 1730-1750 are executed. Otherwise (i.e., the No path of step 1720), steps 1730-1750 are skipped. In step 1730, a target candidate cross-component predictor for the second color block is derived, where a target candidate cross-component model associated with the target candidate cross-component predictor is derived using... Petition 870250094644, dated 10 / 16 / 2025, pp. 99 / 111 92 / 94 reference data for the first corresponding color component and the second corresponding color component for the current block, where the reference data are associated with a reference region comprising a model of the current block or a predefined region indicated using a vector. In step 1740, a final prediction is derived using the target candidate cross-component predictor. In step 1750, the second color block is encoded or decoded using the final prediction. [000280] The flowchart presented aims to illustrate an example of video encoding according to the present invention. A person skilled in the art may modify each step, rearrange the steps, divide a step, or combine steps to implement the present invention without departing from its spirit. In the disclosure, specific syntax and semantics were used to illustrate examples of implementation of embodiments of the present invention. A person skilled in the art may implement the present invention by replacing the syntax and semantics with equivalent syntax and semantics, without departing from the spirit of the present invention. [000281] The above description is presented to enable a person with intermediate skill in the art to practice the present invention as envisioned in the context of a specific application and its requirements. Various modifications to the described embodiments will be apparent to those with skill in the technical field, and the general principles defined herein may be applied to other embodiments. Therefore, the present invention is not intended to be limited to the specific embodiments shown and described, but should be given the broadest scope consistent with the innovative principles and features disclosed herein. In the detailed description above, various specific details are illustrated in order to provide a complete understanding of the present invention. However, persons skilled in the technical field may require further information. Petition 870250094644, dated 10 / 16 / 2025, pages 100 / 111 93 / 94 will understand that the present invention can be put into practice. [000282] The embodiment of the present invention, as described above, can be implemented in various hardware, software, or a combination of both. For example, an embodiment of the present invention may be one or more integrated circuits on a video compression chip or program code integrated into video compression software to perform the process 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 series of functions to be performed by a computer processor, a digital signal processor, a microprocessor, or a field-programmable gate array (FPGA).These processors can be configured to perform specific tasks according to the invention by executing machine-readable software code or firmware code that defines the specific methods incorporated by the invention. The software code or firmware code can be developed in different programming languages ​​and different formats or styles. The software code can also be compiled for different target platforms. However, different code formats, styles, and languages ​​of software code, and other means of configuring the code to perform the tasks according to the invention, will not depart from the spirit and scope of the invention. [000283] The invention can be incorporated into other specific forms without departing from its spirit or essential characteristics. The examples described should 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 the previous description. All modifications that fall within the application and scope of equivalence of the Petition 870250094644, dated 10 / 16 / 2025, pp. 101 / 111 94 / 94 claims should be included in your scope. Petition 870250094644, dated 10 / 16 / 2025, pp. 102 / 111

Claims

1 / 4 CLAIMS 1. A method for video encoding, the method CHARACTERIZED in that it comprises: receiving input data associated with a current block in a current image comprising a first color component and a second color component, wherein the input data comprises pixel data to be encoded on the encoder side or data associated with the current block to be decoded on the decoder side, and wherein the current block comprises a first color block and a second color block; determining whether a target mode is applied to the current block;In response to applying the target mode to the current block: derive a target candidate cross-component predictor for the second color block, wherein a target candidate cross-component model associated with the target candidate cross-component predictor is derived using reference data from the first corresponding color component and the second corresponding color component for the current block, wherein the reference data are associated with a reference region comprising the current block model or a predefined region indicated using a vector; and derive a final prediction using the target candidate cross-component predictor, and encode or decode the second color block using the final prediction.

2. Method, according to claim 1, CHARACTERIZED in that the reference data comprises the compensated prediction of the first corresponding color component and the second corresponding color component for the current block.

3. Method, according to claim 2, CHARACTERIZED in that, when the vector is a block vector or the current block is encoded using the block vector, the offset prediction of the first corresponding color component and the second corresponding color component for the current block is derived using block offset according to the block vector.

4. Method, according to claim 2, CHARACTERIZED in that, when the vector is a motion vector or the current block is encoded using the motion vector, the compensated prediction of the first corresponding color component and the second corresponding color component for the current block is derived using motion compensation according to the motion vector.

5. Method, according to claim 1, CHARACTERIZED in that the reference data are derived using the reconstruction of the first corresponding color component and the second corresponding color component for the reference region.

6. Method according to claim 1, characterized in that the target candidate cross-component model is derived using current prediction samples, model samples, or both.

7. Method, according to claim 1, CHARACTERIZED in that, after the target candidate cross-component model is derived, the target candidate cross-component predictor for the second color block is derived by applying the target candidate cross-component model to the reconstructed first color block.

8. Method, according to claim 1, CHARACTERIZED in that the target candidate cross-component model corresponds to CCLM (Cross-Component Linear Model), MMLM (Multiple-Model CCLM) or CCCM (Cross-Component Convolutional Model).

9. Method, according to claim 1, CHARACTERIZED in that the target candidate cross-component model is selected from a set of candidates comprising multiple candidates or referring to a list of candidates.

10. Method according to claim 1, CHARACTERIZED in that (a) a candidate model associated with the target cross-component candidate model is selected from a set of self-derived candidates consisting of self-derived cross-component candidate models or (b) the candidate model associated with the target cross-component candidate model is selected from using a legacy model or using a self-derived model.

11. Method, according to claim 1, CHARACTERIZED in that a first syntax is signaled or parsed at a TU (Transformation Unit), TB (Transformation Block), CU (Coding Unit), CB (Coding Block) level, or a combination thereof, to indicate how to obtain the target cross-component candidate model for the current block or whether to apply the target mode to the current block.

12. Method, according to claim 1, CHARACTERIZED in that, when an IBC (Intra-Block Copy) related mode or an intermediate prediction mode is used, the block vector-based prediction for the second color block or the intermediate prediction for the second color block is combined with or replaced by the target candidate cross-component predictor.

13. Method, according to claim 1, CHARACTERIZED in that, when an intermediate prediction mode is used for the second color block, one or more hypotheses from one or more candidate cross-component models are combined with one or more hypotheses from the intermediate prediction mode.

14. Method, according to claim 1, CHARACTERIZED in that, if a cross-component-related mode is used to generate prediction samples for the second color block and the current block is encoded using an IBC-related encoding tool or an intermediate, a flag is signaled or analyzed to indicate whether the cross-component-related mode used is inherited from a previously encoded block or derived using a predetermined cross-component mode.

15. Video encoding apparatus, the apparatus CHARACTERIZED by the fact that it comprises one or more electronic components or processors arranged to: receive input data associated with a current block in a current image comprising a first color component and a second color component, wherein the input data comprises pixel data to be encoded on the encoder side or data associated with the current block to be decoded on the decoder side, and wherein the current block comprises a first color block and a second color block; determine whether a target mode is applied to the current block;In response to applying the target mode to the current block: derive a target candidate cross-component predictor for the second color block, wherein a target candidate cross-component model associated with the target candidate cross-component predictor is derived using reference data from the first corresponding color component and the second corresponding color component for the current block, wherein the reference data are associated with a reference region comprising the current block model or a predefined region indicated using a vector; and derive a final prediction using the target candidate cross-component predictor, and encode or decode the second color block using the final prediction. Petition 870250094628, dated 10 / 16 / 2025, p. 12 / 18;

Images