Method, device, and medium for video processing
View Patent ↗Embodiments of the present disclosure provide a method for video processing. The method comprises: applying, during a conversion between a target block of a video and a bitstream of the video, a gradient-based position dependent pre-diction combination to the target block in a coding mode, a gradient of a number of neighboring samples of the target block being used in the gradient-based position dependent prediction combination; and performing the conversion based on the applying.
1 . A method for video processing, comprising:
determining, during a conversion between a target block of a video and a bitstream of the video, based on one or more neighboring prediction or reconstruction samples outside the target block, one or more hypotheses of the target block by using a prediction mode with more than one hypothesis; and
performing the conversion based on the one or more hypotheses, wherein the one or more hypotheses of the target block comprise at least a hypothesis of a target unit of the target block and the one or more neighboring samples to be used to obtain the hypothesis of the target unit is determined based on coded information of the target unit, or wherein the coded information comprises an intra angular mode applied to the target unit, and the one or more neighboring samples comprise at least one neighboring sample to be used for the target unit, the at least one neighboring sample being different from neighboring samples for a further unit of the video other than the target block.
2 . The method of claim 1 , wherein the prediction mode with more than one hypothesis is one of the following:
an overlapped block motion compensation (OBMC) mode,
a multi-hypothesis prediction (MHP) mode,
a geometric partitioning mode (GPM), or
a combination of intra and inter predication (CIIP) mode.
3 . The method of claim 1 , wherein the one or more hypotheses of the target block are determined in an intra prediction mode taking use of neighboring samples, or
wherein the intra prediction mode is one of the following:
an intra planar mode,
an intra direct currency (DC) mode,
an intra angular prediction mode,
an intra derived mode (DM), or
an intra linear model (LM).
4 . The method of claim 1 , wherein the one or more hypotheses of the target block are determined in an inter prediction mode taking use of neighboring prediction or reconstruction samples, or
wherein the inter prediction mode comprises at least one of an inter local illumination compensation (LIC) mode, an inter overlapped block motion compensation (OBMC) mode, an inter template matching mode or an inter filtering mode using neighboring samples.
5 . The method of claim 1 , wherein the target unit comprises a subblock or a partition of the target block.
6 . The method of claim 1 , wherein the target unit comprises a hypothesis of a plurality of hypotheses of the target block.
7 . The method of claim 1 , wherein
the coded information of the targe unit comprises coded information of the hypothesis of the target unit, the coded information of the hypothesis indicating that a prediction direction is from left and/or above, and
the one or more neighboring samples comprise more than one left and/or above neighboring sample grouped together to construct a template for prediction samples of the hypothesis.
8 . The method of claim 7 , wherein the hypothesis of the target unit is determined in an intra mode.
9 . The method of claim 1 , wherein whether the one or more neighboring samples are used to obtain the hypothesis of the target unit is determined based on coded information of the target unit.
10 . The method of claim 9 , wherein the coded information of the target unit comprises at least one of the following:
availability of the one or more neighboring samples, or
at least one of a partitioning shape, a partition angle or a direction associated with the target unit.
11 . The method of claim 10 , wherein the target unit comprises a partition or subblock of the target block, a hypothesis of a plurality of hypotheses of the target block, or the target block.
12 . The method of claim 10 , wherein
the target unit comprises a partition or subblock of the target block or a hypothesis of a plurality of hypotheses of the target block in a geometric partitioning merge mode, and
the one or more neighboring samples comprise left and/or above neighboring samples.
13 . The method of claim 12 , wherein
the coded information of the target unit comprises a partitioning shape of the geometric partitioning merge mode, and
wherein whether the left and/or above neighboring samples are used for prediction samples of the hypothesis is determined based on the partitioning shape of the geometric partitioning merge mode, or
wherein the partitioning shape of the geometric partitioning merge mode is indicated by an index for merge GPM partition.
14 . The method of claim 12 , wherein
the coded information of the target unit comprises a partition angle of the geometric partitioning merge mode, and
wherein whether the left and/or above neighboring samples are used for prediction samples of the hypothesis is determined based on the partition angle of the geometric partitioning merge mode, or
wherein the partition angle of the geometric partitioning merge mode is derived from a partitioning shape of the geometric partitioning merge mode, or
wherein the partition angle of the geometric partitioning merge mode is included indicated by an index for angle.
15 . The method of claim 14 , wherein
the coded information of the target unit comprises a partition distance of the geometric partitioning merge mode, and
wherein whether the left and/or above neighboring samples are used for prediction samples of the hypothesis is determined based on the partition distance of the geometric partitioning merge mode.
16 . The method of claim 1 , wherein the conversion includes encoding the target block into the bitstream, or
wherein the conversion includes decoding the target block from the bitstream.
17 . The method of claim 1 , further comprising:
storing the bitstream in a non-transitory computer-readable recording medium.
18 . An apparatus for processing video data comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to perform a method for video processing, comprising:
determining, during a conversion between a target block of a video and a bitstream of the video, based on one or more neighboring prediction or reconstruction samples outside the target block, one or more hypotheses of the target block by using a prediction mode with more than one hypothesis; and
performing the conversion based on the one or more hypotheses, wherein the one or more hypotheses of the target block comprise at least a hypothesis of a target unit of the target block and the one or more neighboring samples to be used to obtain the hypothesis of the target unit is determined based on coded information of the target unit, or wherein the coded information comprises an intra angular mode applied to the target unit, and the one or more neighboring samples comprise at least one neighboring sample to be used for the target unit, the at least one neighboring sample being different from neighboring samples for a further unit of the video other than the target block.
19 . A non-transitory computer-readable storage medium storing instructions that cause a processor to perform a method for video processing, comprising:
determining, during a conversion between a target block of a video and a bitstream of the video, based on one or more neighboring prediction or reconstruction samples outside the target block, one or more hypotheses of the target block by using a prediction mode with more than one hypothesis; and
performing the conversion based on the one or more hypotheses, wherein the one or more hypotheses of the target block comprise at least a hypothesis of a target unit of the target block and the one or more neighboring samples to be used to obtain the hypothesis of the target unit is determined based on coded information of the target unit, or wherein the coded information comprises an intra angular mode applied to the target unit, and the one or more neighboring samples comprise at least one neighboring sample to be used for the target unit, the at least one neighboring sample being different from neighboring samples for a further unit of the video other than the target block.