Affine based merge with MVD
AssigneesBeijing Bytedance Network Technology Co., Ltd., Bytedance Inc.
InventorsKai Zhang, Li Zhang, Hongbin Liu, Jizheng Xu, Yue Wang
Life of this patent
Estimating the term…
Abstract
Disclosed and systems, methods and devices for video processing. An exemplary method for video processing includes: determining for a current video block, an indication which indicates that an affine merge with motion vector difference (MMVD) mode is applied to the current video block; performing a conversion between the current video block and a bitstream representation of a current video block based on the indication, the indication being activated based on the usage of at least one merge list for the current video block.
Claims18
A method for video processing, comprising:
- determining that an affine mode is enabled for a first block of a video;
- in response to the affine mode being enabled for the first block, determining whether a first ultimate motion vector expression mode related with the affine mode is applied to the first block;
- in response to the first ultimate motion vector expression mode related with the affine mode being applied to the first block, determining a first ultimate motion vector expression mode information for the first block, wherein the first ultimate motion vector expression mode information comprises a first distance index and a first direction information, the first distance index indicates an offset of a refined motion information respective to an original motion information of the first block and the first direction information indicates a direction of the offset;
- and
- performing a conversion between the first block and a bitstream of the video using the first ultimate motion vector expression mode information,
- wherein a second ultimate motion vector expression mode related with a non-affine mode is applied to a second block coded with the non-affine mode, wherein the second ultimate motion vector expression mode corresponds to a second ultimate motion vector expression mode information, and wherein the second ultimate motion vector expression mode information comprises a second distance index indicating an offset of a refined motion information respective to an original motion information of the second block and a second direction information indicating a direction of the offset.
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:
- determine that an affine mode is enabled for a first block of a video;
- in response to the affine mode being enabled for the first block, determine whether a first ultimate motion vector expression mode related with the affine mode is applied to the first block;
- in response to the first ultimate motion vector expression mode related with the affine mode being applied to the first block, determine a first ultimate motion vector expression mode information for the first block, wherein the first ultimate motion vector expression mode information comprises a first distance index and a first direction information, the first distance index indicates an offset of a refined motion information respective to an original motion information of the first block and the first direction information indicates a direction of the offset;
- and
- perform a conversion between the first block and a bitstream of the video using the first ultimate motion vector expression mode information,
- wherein a second ultimate motion vector expression mode related with a non-affine mode is applied to a second block coded with the non-affine mode, wherein the second ultimate motion vector expression mode corresponds to a second ultimate motion vector expression mode information, wherein the second ultimate motion vector expression mode information comprises a second distance index indicating an offset of a refined motion information respective to an original motion information of the second block and a second direction information indicating a direction of the offset.
A non-transitory computer-readable storage medium storing instructions that cause a processor to:
- determine that an affine mode is enabled for a first block of a video;
- in response to the affine mode being enabled for the first block, determine whether a first ultimate motion vector expression mode related with the affine mode is applied to the first block;
- in response to the first ultimate motion vector expression mode related with the affine mode being applied to the first block, determine a first ultimate motion vector expression mode information for the first block, wherein the first ultimate motion vector expression mode information comprises a first distance index and a first direction information, the first distance index indicates an offset of a refined motion information respective to an original motion information of the first block and the first direction information indicates a direction of the offset;
- and
- perform a conversion between the first block and a bitstream of the video using the first ultimate motion vector expression mode information;
- wherein a second ultimate motion vector expression mode related with a non-affine mode is applied to a second block coded with the non-affine mode, wherein the second ultimate motion vector expression mode corresponds to a second ultimate motion vector expression mode information, wherein the second ultimate motion vector expression mode information comprises a second distance index indicating an offset of a refined motion information respective to an original motion information of the second block and a second direction information indicating a direction of the offset.
A method for storing a bitstream of a video, comprising:
- determining that an affine mode is enabled for a first block of the video;
- in response to the affine mode being enabled for the first block, determining whether a first ultimate motion vector expression mode related with the affine mode is applied to the first block;
- in response to the first ultimate motion vector expression mode related with the affine mode being applied to the first block, determining a first ultimate motion vector expression mode information for the first block, wherein the first ultimate motion vector expression mode information comprises a first distance index and a first direction information, the first distance index indicates an offset of a refined motion information respective to an original motion information of the first block and the first direction information indicates a direction of the offset;
- generating the bitstream using the first ultimate motion vector expression mode information;
- and
- storing the bitstream in a non-transitory computer-readable recording medium,
- wherein a second ultimate motion vector expression mode related with a non-affine mode is applied to a second block coded with the non-affine mode, wherein the second ultimate motion vector expression mode corresponds to a second ultimate motion vector expression mode information, wherein the second ultimate motion vector expression mode information comprises a second distance index indicating an offset of a refined motion information respective to an original motion information of the second block and a second direction information indicating a direction of the offset.
Dependencies are inferred from the claim text.