Subblock Motion Vector Coding for Precise Video Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional subblock motion modes in video encoding and decoding do not allow for customized assignment of motion vectors, leading to suboptimal encoding and decoding efficiency.
Innovation Solution
Derive and utilize subblock motion vector predictors (MVPs) and differences (MVDs) to improve the accuracy and coding efficiency of motion vectors for video blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional subblock motion modes are used, then the encoding and decoding process is simple, but the motion vector accuracy and coding efficiency are suboptimal
Solution Approach 1:
The current block is divided into multiple subblocks, and each subblock is assigned its own motion vector independently. This segmentation allows for more precise motion representation in different regions of the block, improving motion vector accuracy while maintaining manageable encoding complexity through systematic processing.
Solution Approach 2:
Different motion vectors are assigned to different subblocks based on local motion characteristics. This local quality approach allows each subblock to have customized motion parameters tailored to its specific motion behavior, improving overall coding efficiency without requiring complex global optimization.
2Reliability
If conventional motion vector assignment is used, then the encoding process is fast, but the video quality and encoding accuracy are reduced
Solution Approach 1:
Motion vector predictors are derived in advance from neighboring blocks and reference frames before the actual encoding process. This preliminary action provides accurate starting points for motion vector calculation, improving encoding accuracy while reducing the computational burden during the main encoding phase, thus maintaining encoding speed.
Solution Approach 2:
The system uses previously encoded motion information from neighboring blocks and reference frames to generate motion vector predictors automatically. This self-service mechanism improves encoding accuracy by leveraging existing data without requiring additional complex computations or external information, maintaining productivity.
Data Source
AI summary
The various embodiments described herein include methods and systems for encoding and decoding video. In one aspect, a method includes receiving a video bitstream comprising a plurality of blocks, including a first block, where the first block includes a plurality of subblocks. The method also includes determining, based on a syntax element value in the video bitstream, that the first block is predicted in a subblock mode. The method further includes, in accordance with the first block being predicted in the subblock mode: deriving a set of motion vector predictors (MVPs) corresponding to a plurality of subblocks of the first block; deriving a set of motion vector differences (MVDs) for the plurality of subblocks; determining a set of motion vectors for the plurality of subblocks by applying the set of MVDs to the set of MVPs; and reconstructing the first block using the set of motion vectors.