Abstract: A method and apparatus of motion compensation using the bi-directional optical flow (BIO) techniques are disclosed. According to one method, the use of BIO is extended to general bi-prediction motion compensation by including the case that two reference pictures correspond to two previously coded pictures. According to another method, the use of BIO is adaptively applied depending on the linearity of the two motion vectors associated with the two reference blocks or depending on block size of the current block. According to yet another method, the refined motion vectors by compensating the original motion vectors with the respective x-offset values and y-offset values are stored in a motion-vector buffer for motion vector prediction of one or more following blocks.

Abstract: A method and apparatus of video coding incorporating Deep Neural Network are disclosed. A target signal is processed using DNN (Deep Neural Network), where the target signal provided to DNN input corresponds to the reconstructed residual, output from the prediction process, the reconstruction process, one or more filtering processes, or a combination of them. The output data from DNN output is provided for the encoding process or the decoding process. The DNN can be used to restore pixel values of the target signal or to predict a sign of one or more residual pixels between the target signal and an original signal. An absolute value of one or more residual pixels can be signalled in the video bitstream and used with the sign to reduce residual error of the target signal.

Abstract: Method and apparatus of video coding using decoder derived motion information based on bilateral matching or template matching are disclosed. According to one method, an initial motion vector (MV) index is signalled in a video bitstream at an encoder side or determined from the video bitstream at a decoder side. A selected MV is then derived using bilateral matching, template matching or both to refine an initial MV associated with the initial MV index. In another method, when both MVs for list 0 and list 1 exist in template matching, the smallest-cost MV between the two MVs may be used for uni-prediction template matching if the cost is lower than the bi-prediction template matching. According to yet another method, the refinement of the MV search is dependent on the block size. According to yet another method, merge candidate MV pair is always used for bilateral matching or template matching.

Abstract: Methods for decoding of a video bitstream by a video decoding circuit are provided. In one implementation, a method receives coded data for a 2N×2N coding unit (CU) from the video bitstream, selects one or more first codewords according to whether asymmetric motion partition is disabled or enabled when a size of said 2N×2N CU is not equal to a smallest CU size, wherein none of the first codewords corresponds to INTER N×N partition, selects one or more second codewords when the size of said 2N×2N CU is equal to the smallest CU size, wherein none of the second codewords corresponds to the INTER N×N partition when N is 4, determines a CU structure for said 2N×2N CU from the video bitstream using said one or more first codewords or said one or more second codewords, and decodes the video bitstream using the CU structure.

Abstract: A method and apparatus derive a motion vector predictor (MVP) for a current block in an Inter, Merge, or Skip mode. Embodiments according to the present invention determine redundant MVP candidates according to a non-MV-value based criterion. The redundant MVP candidates are then removed from the MVP candidate set. In other embodiments according to the present invention, motion IDs are assigned to MVP candidates to follow the trail of motion vectors associated with the MVP candidate. An MVP candidate having a same motion ID as a previous MVP is redundant and can be removed from the MVP candidate set. In yet another embodiment, redundant MVP candidates correspond to one or more of the MVP candidates that cause the second 2N×N or N×2N PU to be merged into a 2N×2N PU are removed from the MVP candidate set.

Abstract: Video processing methods and apparatuses for candidate set determination for a current block partitioned from a parent block by quad-tree splitting comprise receiving input data of a current block, determining a candidate set for the current block by prohibiting a spatial candidate derived from any of neighboring blocks partitioned from the same parent block or determining the candidate set for the current block by conducting a pruning process if all the neighboring blocks are coded in Inter prediction and motion information of the neighboring blocks are the same, and encoding or decoding the current block based on the candidate set by selecting one final candidate from the candidate set. The pruning process comprises scanning the candidate set to determine if any candidate equals to the spatial candidate derived from the neighboring blocks, and removing the candidate equals to the spatial candidate from the candidate set.

Abstract: An image processing method applied to an image processing system. The image processing method comprises: (a) computing an image intensity distribution of an input image; (b) performing atmospheric light estimation to the input image; (c) performing transmission estimation according to a result of the step (a) to the input image, to generate a transmission estimation parameter; and (d) recovering scene radiance of the input image according to a result generated by the step (b) and the transmission estimation parameter. At least one of the steps (a)-(c) are performed to data corresponding to only partial pixels of the input image.

Abstract: Aspects of the disclosure provide a method for video coding in merge mode or skip mode. The method can include receiving a prediction block (PB) of a picture, determining number and positions of merge candidates of the PB according to a size and/or a shape of the PB, and constructing a candidate list including motion data of a subset of the merge candidate positions.

Abstract: A method for performing parallel coding with ordered entropy slices includes: providing a plurality of entropy slices to a plurality of processing elements, wherein each entropy slice includes a plurality of blocks; initializing CABAC states of a current entropy slice as the CABAC states of a previous entropy slice after processing DB blocks of the previous entropy slice. DB is a positive integer.

Abstract: Aspects of the disclosure provide a method for denoising a reconstructed picture in a video coding system. The method can include providing two or more candidate non-local denoising technologies, selecting a target non-local denoising technology from the two or more candidate non-local denoising technologies for a patch group, and denoising the patch group of the reconstructed picture with the target non-local denoising technology. Besides, two parallel processing methods are provided for forming patch groups with predictor-based searching algorithms.

Abstract: Method and apparatus of video coding using generalized bi-prediction are disclosed. According to one method, the generalized bi-prediction is extended for Decoder-Side Motion Vector Refinement (DMVR), where unequal weighting factor pair is allowed to form the template of the current block or the final predictor. In another method, the generalized bi-prediction is extended to pattern-based MV derivation (PMVD), where unequal weighting factor pair is allowed to combined reference blocks pointed by motion vectors derived based on PMVD. In yet another method, the generalized bi-prediction is used for Merge mode. When the final Merge MVP selected corresponds to bi-prediction using an unequal weighting factor pair, a new weighting factor pair is derived using neighboring reconstructed pixels of the current block and corresponding motion compensated pixels. In yet another method, the size of a set of weighting factor pairs is dependent on block size.

Abstract: A method and apparatus for coding video data using Inter prediction mode or Merge mode in a video coding system are disclosed, where the video data is configured into a Base Layer (BL) and an Enhancement Layer (EL), and the EL has higher spatial resolution or better video quality than the BL. In one embodiment, at least one information piece of motion information associated with one or more BL blocks in the BL is identified. A motion vector prediction (MVP) candidate list or a Merge candidate list for the selected block in the EL is then determined, where said at least one information piece associated with said one or more BL blocks in the BL is included in the MVP candidate list or the Merge candidate list. The input data associated with the selected block is coded or decoded using the MVP candidate list or the Merge candidate list.

Abstract: A method and apparatus for deblocking reconstructed video in a video coding system. The method receives reconstructed pixel data, wherein the reconstructed pixel data is configured into LCUs (largest coding units) and each LCU is divided into blocks. The method then identifies horizontal block edges of the blocks in the LCUs, applies first vertical filtering to the reconstructed pixel data corresponding to one or more first lines above a first horizontal block edge in the first edge group, and applies second vertical filtering to the reconstructed pixel data corresponding to one or more second lines above a second horizontal block edge in the second edge group, wherein a first number of said one or more first lines is smaller than a second number of said one or more second lines.

Abstract: A method and apparatus for processing of coded video using in-loop processing are disclosed. The method operates by receiving input data to said in-loop processing, wherein the input data corresponds to reconstructed coding units of the picture; configuring the input data into multiple filter units; selecting a filter from a candidate filter set comprising at least two candidate filters associated with said in-loop processing for one of said multiple filter units; applying said in-loop processing to said one of said multiple filter units using the selected filter to generate a processed filter unit, wherein when said one of said multiple filter units comprises at least two reconstructed coding units, the selected filter is applied to all of said at least two reconstructed coding units; and providing processed video data comprising the processed filter unit. The apparatus provide circuits to carryout the operations of the method.

Abstract: In order to overcome the issue caused by a decoded block vector (BV) pointing to a reference block overlapping with an unavailable area, various methods are disclosed. According to one method, if the reference block overlaps with an unavailable area, the reference pixels in the unavailable area are generated for IntraBC prediction of the current block by padding from neighbouring available pixels. The padding can be done in the horizontal direction and then the vertical direction. The padding may also done in the vertical direction first and then horizontal direction. In another method, if the reference block overlaps with an unavailable area, the reference pixels in the unavailable area are generated for IntraBC prediction of the current block by using previous decoded pixels in the unavailable area. A pre-defined value may also be used for the unavailable area.

Abstract: Methods of palette coding to reduce the required coding process are disclosed. According to one method, smaller blocks are derived from a large block. The histogram of the large block is derived based on the histograms of smaller blocks in the large block. According to another method, one or more palette tables are derived based on multiple blocks. One palette table is used for each of the multiple blocks. According to yet another method, index map transpose is performed in the parsing stage according to the transpose flag of the index map. Accordingly, a buffer to store the transpose flags can be saved. According to still yet another method, the palette predictor update is performed using an index mapping table to avoid the need for shuffling the contents of the palette predictor stored in a palette buffer.

Abstract: A method and apparatus of line buffer reduction for context adaptive entropy processing are disclosed. The context formation for context adaptive entropy processing depends on block information associated with one or more neighboring blocks. When a first neighboring block is on an upper side of a horizontal region boundary or on a left side of a vertical region boundary of the region, the block information is replaced by replacement block information to reduce or remove line buffer requirement for storing the block information of neighboring blocks on the other side of the region boundaries from the current block. The context adaptive entropy processing is CABAC encoding, CABAC decoding, CAVLC encoding, or CAVLC decoding.

Abstract: A method and apparatus of entropy coding for a video encoder or decoder using multiple-table based Context-Based Adaptive Binary Arithmetic Coder (CABAC) are disclosed. In one embodiment, a current bin of a binary data of a current coding symbol is encoded or decoded according to a probability of a binary value of the current bin and the probability of the binary value is updated according to the binary value of the current bin for a next bin by using multiple-parameter probability models. Each multiple-parameter probability model is updated using at least one lookup table with the individual set of probability state as a table index to access contents of said at least one lookup table. In another embodiment, the range update is calculated for a range interval based on middle value of the range interval.

Abstract: A depth image processing method and a depth image processing system are provided. The depth image processing method includes: capturing a first image and a second image; performing a feature comparison to acquire a plurality of feature pairs between the first image and the second image, wherein each of the feature pairs includes a feature in the first image and a corresponding feature in the second image; computing disparities of the feature pairs; computing a depth image through the first image and the second image when the disparities of the feature pairs are all smaller than a disparity threshold.

Abstract: A method and apparatus for Intra prediction of non-square chroma blocks are disclosed. The system uses a block partitioning structure for partitioning a chroma block into one or more final sub-blocks including at least one non-square final sub-block using block partition recursively. The block partition either splits each given block into intermediate sub-blocks using one or more split modes including one binary split mode or applies no splitting to each given block. The block is treated as an initial given block and each intermediate sub-block is treated as one given block in a next recursion. Any given block that is not further split becomes a final sub-block. After the final sub-blocks are determined, non-square Intra prediction is applied to each non-square final sub-block.