摘要:
A video coder performs a padding operation that processes a set of border pixels according to an order. The order starts at a bottom-left border pixel and proceeds through the border pixels sequentially to a top-right border pixel. When the padding operation processes an unavailable border pixel, the padding operation predicts a value of the unavailable border pixel based on a value of a border pixel previously processed by the padding operation. The video coder may generate an intra-predicted video block based on the border pixels.
摘要:
Aspects of this disclosure relate to a method of coding video data. In an example, the method includes determining a first residual quadtree (RQT) depth at which to apply a first transform to luma information associated with a block of video data, wherein the RQT represents a manner in which transforms are applied to luma information and chroma information. The method also includes determining a second RQT depth at which to apply a second transform to the chroma information associated with the block of video data, wherein the second RQT depth is different than the first RQT depth. The method also includes coding the luma information at the first RQT depth and the chroma information at the second RQT depth.
摘要:
In one example, this disclosure describes a method of codeword adaptation for variable length coding. The method comprises applying a first codeword adaptation scheme to a first group of codewords of a variable length coding (VLC) table to change a mapping of codewords to events in the VLC table; and applying a second codeword adaptation scheme to a second group of codewords of the VLC table to change the mapping of the codewords to the events in the VLC table.
摘要:
This disclosure describes techniques for coding transform coefficients for a block of video data. According to one aspect of this disclosure, a coder (e.g., an encoder or decoder) may map between a code number cn and a level_ID value and an run value based on a structured mapping. According to other aspects of this disclosure, the coder may map between a code number cn and a level_ID value and an run value for the current transform coefficient using a first technique or a second technique based on a coded block type of a block of video data being coded. For example, if the coded block type is a first coded block type, the coder may use a structured mapping. However, if the coded block type is a second coded block type different than the first coded block type, the coder may access one or more mapping tables stored in memory to perform the mapping.
摘要:
In one example, this disclosure describes method of coding video data. The method comprises coding a block of video data as one or more luminance blocks of transform coefficients and one or more chrominance blocks of transform coefficients, and coding a coded block pattern (CBP) for the block of video data. The CBP comprises syntax information that identifies whether non-zero data is included in each of the luminance blocks and each of the chrominance blocks. Coding the CBP includes selecting one or more variable length coding (VLC) tables based on a transform size used in performing one or more transforms on the one or more luminance blocks.
摘要:
A method and system for entropy coding can comprise, in response to detecting a first symbol combination comprising first run information indicating a first number of contiguous zero coefficients is greater than a cut-off-run value, assigning a first codeword to a first symbol combination, wherein the first codeword comprises an escape code from a first-level VLC table; and in response to a second symbol combination comprising second run information indicating a second number of contiguous zero coefficients is less than or equal to the cut-off-run value, assigning a second codeword to the second symbol combination, wherein the second codeword is from the first-level VLC table. The system and method can further comprise collecting coding statistics for a set of candidate symbol combinations and adjusting a mapping between codewords of the first-level VLC table and a subset of the set of candidate symbol combinations based on the coding statistics.
摘要:
In one example, an apparatus includes a video encoder configured to partition a block of video data into a first partition and a second partition using a geometric motion partition line, calculate a prediction value of a pixel in a transition region of the block using a filter that applies a value for at least one neighboring pixel from the first partition and a value for at least one neighboring pixel from the second partition, calculate a residual value of the pixel in the transition region of the block based on the prediction value of the pixel in the transition region, and output the residual value of the pixel. In one example, a video decoder may use a similar filter to decode an the encoded block after receiving the residual value for the encoded block, and using a definition of the geometric motion partition line.
摘要:
A system and method for providing improved FGS identification in scalable video coding. According to the present invention, each FGS enhancement layer is assigned a unique dependency identifier and contains only FGS enhancement information. For subsequent enhancement layers, the base dependency identifier for the subsequent enhancement layers will point to either a base-quality layer or an FGS enhancement layer.
摘要:
The present invention involves the use of the FRExt approach for FGS. According to the present invention, an 8×8 data block is de-interleaved and processed as individual 4×4 data blocks, with an additional end-of-8×8-block (EO8B) marker indicating that no more coefficients remain in any of the de-interleaved 4×4 data blocks. The EO8B symbol may be a binary flag. The invention also uses a longer codeword for the EO8B symbol, conveying information about which de-interleaved blocks contain additional coefficients.
摘要:
Methods, devices, and computer code products for encoding and decoding a video signal including conditional encoding and decoding of a residual prediction flag for an enhancement layer only of all base layers are discrete layers. If some base layers are not discrete, the residual prediction flag is always encoded and decoded. Encoding and decoding the residual prediction flag can include using contexts which depend on whether the reconstructed prediction residual of the discrete base layers is zero or not.