US-12621481-B2 - Selective application of decoder side refining tools

Abstract

One example method of video processing includes implementing, by a processor, a decoder-side motion vector derivation (DMVD) scheme for motion vector refinement during a conversion between a current video block and a bitstream representation of the current video block by deriving parameters based on a deriving rule. The conversion may include compressing the current video block into the bitstream representation or uncompressing the bitstream representation into pixel values of the current video block.

Inventors

• Hongbin Liu
• Li Zhang
• Kai Zhang
• Jizheng Xu
• Yue Wang

Assignees

• BEIJING BYTEDANCE NETWORK TECHNOLOGY CO., LTD.
• BYTEDANCE INC.

Dates

Publication Date

20260505

Application Date

20240703

Priority Date

20190214

Claims (19)

1. 1 . A method of processing video data, comprising: obtaining a first refined motion information for a first current video block of a video by implementing a decoder-side motion vector derivation (DMVD) scheme for the first current video block, wherein a symmetric motion vector difference (SMVD) mode is enabled for the first current video block; and performing a first conversion between the first current video block and a bitstream of the video using the first refined motion information, wherein the DMVD scheme includes a decoder-side motion vector refinement (DMVR) scheme, and/or a bi-directional optical flow (BDOF) scheme, and wherein the method further comprises determining, based on a field in the bitstream, whether to enable or disable the DMVD scheme for the first current video block in the SMVD mode before obtaining the first refined motion information for the first current video block.
2. 2 . The method of claim 1 , wherein the bitstream includes a motion vector difference (MVD) for the first refined motion information, and wherein the MVD is decoded according to the SMVD mode and is further refined before being used to decode the first current video block.
3. 3 . The method of claim 1 , wherein the bitstream does not include a signaling for a motion vector difference (MVD) for the first refined motion information in the SMVD mode, and wherein the MVD is generated using the DMVD scheme.
4. 4 . The method of claim 1 , wherein a motion vector difference (MVD) is not signaled in the bitstream in response to the DMVD scheme being enabled.
5. 5 . The method of claim 1 , wherein the field indicating whether the DMVD scheme is enabled or disabled is present in the bitstream for one or more motion vector (MV) precisions or motion vector difference (MVD) precisions, or wherein the field indicating whether the DMVD scheme is enabled or disabled is signaled in the bitstream in response to a first picture order count distance (PocDis0) representing a first distance from a first reference picture (Ref0) of the first current video block to a current picture being equal to a second picture order count distance (PocDis1) representing a second distance from the current picture to a second reference picture (Ref1) of the first current video block, and wherein one of the first and second reference picture (Ref0) precedes the current picture and the other of the first and second reference picture (Ref1) follows the current picture in a display order.
6. 6 . The method of claim 5 , wherein the one or more MV precisions or MVD precisions include 1-pel and/or 4-pel precision.
7. 7 . The method of claim 1 , wherein the DMVD scheme is enabled in SMVD mode based on one of: a coding information of the first current video block and/or of one or more neighboring blocks; a block dimension of the first current video block; information related to reference pictures for the first current video block, wherein information related to the reference pictures includes picture order count (POC) information; or a signaling for a motion vector difference (MVD) information in the bitstream.
8. 8 . The method of claim 1 , further comprising: obtaining a second refined motion information for a second current video block of the video by implementing the DMVD scheme based on at least a weight parameter, wherein the weight parameter is applied to prediction blocks in a generation process for a final prediction block for the second current video block; and performing a second conversion between the second current video block and the bitstream of the video at least using the second refined motion information and the weight parameter.
9. 9 . The method of claim 8 , wherein a field in the bitstream is indicative of the weight parameter.
10. 10 . The method of claim 8 , wherein an indication of the weight parameter is signaled using a Bi-prediction with Coding unit Weights (BCW) technique or using a weighted prediction technique.
11. 11 . The method of claim 8 , wherein the DMVD scheme is implemented by applying the weight parameter to prediction blocks of the second current video block.
12. 12 . The method of claim 11 , wherein the second conversion includes calculating a predictive cost function for the second current video block by first applying the weight parameter according to a Bi-prediction with Coding unit Weights (BCW) index to the prediction blocks and then calculating the predictive cost function.
13. 13 . The method of claim 11 , wherein the second conversion includes calculating a predictive cost function for the second current video block by first applying the weight parameter according to a weighted prediction scheme to the prediction blocks and then calculating the predictive cost function.
14. 14 . The method of claim 12 , wherein the predictive cost function is selected from one of: a gradient function, a sum of absolute difference (SAD) cost function, or a mean-removed sum of absolute difference (MR-SAD) cost function.
15. 15 . The method of claim 8 , wherein a reshaping process is disabled for prediction blocks generated with the second refined motion information for the second current video block.
16. 16 . The method of claim 1 , wherein the first conversion includes encoding the first current video block into the bitstream.
17. 17 . The method of claim 1 , wherein the first conversion includes decoding the first current video block from the bitstream.
18. 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: obtain a first refined motion information for a first current video block of a video by implementing a decoder-side motion vector derivation (DMVD) scheme for the first current video block wherein a symmetric motion vector difference (SMVD) mode is enabled for the first current video block; and perform a first conversion between the first current video block and a bitstream of the video using the first refined motion information, wherein the DMVD scheme includes a decoder-side motion vector refinement (DMVR) scheme, and/or a bi-directional optical flow (BDOF) scheme, and wherein the instructions further cause the processor to determine, based on a field in the bitstream, whether to enable or disable the DMVD scheme for the first current video block in the SMVD mode before obtaining the first refined motion information for the first current video block.
19. 19 . A non-transitory computer-readable recording medium storing a bitstream which is generated by a method performed by a video processing apparatus, wherein the method comprises: obtaining a first refined motion information for a first current video block of a video by implementing a decoder-side motion vector derivation (DMVD) scheme for the first current video block wherein a symmetric motion vector difference (SMVD) mode is enabled for the first current video block; and generating the bitstream using the first refined motion information, wherein the DMVD scheme includes a decoder-side motion vector refinement (DMVR) scheme, and/or a bi-directional optical flow (BDOF) scheme, and wherein the method further comprises determining, based on a field in the bitstream, whether to enable or disable the DMVD scheme for the first current video block in the SMVD mode before obtaining the first refined motion information for the first current video block.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application is a continuation of U.S. application Ser. No. 17/401,438, filed on Aug. 13, 2021, which is a continuation of International Application No. PCT/CN2020/075232, filed on Feb. 14, 2020, which claims the priority to and benefits of International Patent Application No. PCT/CN2019/075068, filed on Feb. 14, 2019, and International Patent Application No. PCT/CN2019/082585, filed on Apr. 13, 2019. The entire disclosures of the aforementioned applications are incorporated by reference as part of the disclosure of this application. TECHNICAL FIELD The present disclosure relates to video and image coding and decoding. BACKGROUND Digital video accounts for the largest bandwidth use on the internet and other digital communication networks. As the number of connected user devices capable of receiving and displaying video increases, it is expected that the bandwidth demand for digital video usage will continue to grow. SUMMARY The present disclosure discloses video coding tools that, in one example aspect, improve coding efficiency of current coding tools related to ultimate motion vector expression or generalized bi-prediction. A first example video processing method comprises obtaining a refined motion information for a current video block of a video by implementing a decoder-side motion vector derivation (DMVD) scheme based on at least a weight parameter, where the weight parameter is applied to prediction blocks in a generation process for a final prediction block for the current video block, and performing a conversion between the current video block and a bitstream representation of the video at least using the refined motion information and the weight parameter. A second example video processing method comprises determining that use of a decoder-side motion vector derivation (DMVD) scheme is disabled for a conversion between a current video block of a video and a coded representation of the video due to use of a coding tool for the current video block, and performing the conversion between the current video block and a bitstream representation of the video based on the determining, where the coding tool includes applying unequal weighting factors to prediction blocks of the current video block. A third example video processing method comprises determining, based on picture order count (POC) values of one or more reference pictures of a current video block of a video and a POC value of a current picture containing the current video block, whether to enable or disable one or more decoder-side motion vector derivation (DMVD) schemes for the current video block, and performing, according to the determining, a conversion between the current video block and a bitstream representation of the video. A fourth example video processing method comprises obtaining a refined motion information for a current video block of a video by implementing a decoder-side motion vector derivation (DMVD) scheme for the current video block where a symmetric motion vector difference (SMVD) mode is enabled for the current video block, and performing a conversion between the current video block and a bitstream representation of the video using the refined motion information. A fifth example video processing method comprises determining, based on a field in a bitstream representation for a video including a current video block, whether to enable or disable a decoder-side motion vector derivation (DMVD) scheme for the current video block, where a symmetric motion vector difference (SMVD) mode is enabled for the current video block, obtaining, after the determining that the DMVD scheme is enabled, a refined motion information for the current video block by implementing the DMVD scheme for the current video block, and performing a conversion between the current video block and a bitstream representation of the video using the refined motion information. A sixth example video processing method comprises determining, based on a rule that uses a block dimension of a current video block of a video, whether a plurality of decoder-side motion vector derivation (DMVD) schemes are enabled or disabled for a conversion between the current video block and a bitstream representation of the video, and performing the conversion based on the determination. A seventh example video processing method comprises determining whether to perform a plurality of decoder-side motion vector derivation (DMVD) schemes at a sub-block level or a block level for a current video block of a video, obtaining, after the determining that the plurality of DMVD schemes are performed at a sub-block level, a refined motion information for the current video block by implementing the plurality of DMVD schemes at a same sub-block level for the current video block, and performing a conversion between the current video block and a bitstream representation of the video using the refined motion information. An eighth example video processing