Reference Motion Vector Bank Updates for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing reference motion vector candidate bank updating process in video coding does not consider motion vector candidates from the same superblock, leading to suboptimal results due to the failure to consider other possible directions and blocks.
Innovation Solution
A method for decoding video data that involves retrieving motion vector predictors from a reference motion vector candidate bank, determining the motion vector for the current block, updating the reference motion vector candidate bank with the new motion vector, and decoding the current block based on the determined motion vector. This process considers motion vectors from the same superblock for more efficient prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the reference motion vector candidate bank is updated only with motion vectors from different superblocks, then the updating process is simple, but the coding performance is suboptimal due to limited candidate diversity
Solution Approach 1:
The patent segments the motion vector candidate selection process by block type. Intra-block motion vectors are excluded from updating the reference bank, while inter-block motion vectors are included. This segmentation allows the system to maintain simple updating logic while improving candidate diversity and coding performance by selectively incorporating relevant motion vectors.
2Measurement precision
If motion vectors from the same superblock are considered for prediction, then the prediction accuracy improves, but the latency increases due to additional processing
Solution Approach 1:
The patent performs preliminary action by pre-establishing the reference motion vector candidate bank during the decoding process before final prediction. Motion vectors from already-decoded blocks within the same superblock are pre-collected and stored in the reference bank, making them readily available for prediction without adding latency during the actual prediction phase. This preliminary preparation enables accurate same-superblock prediction while maintaining processing efficiency.
Data Source
AI summary
Method, device, and non-transitory storage medium for decoding video data are provided. One or more vector predictors associated with a current block from a reference motion vector candidate bank may be retrieved, the one or more retrieved motion vector predictors comprising at least one or more motion vectors associated with one or more already decoded blocks, and the one or more already decoded blocks belonging to a same superblock as the current block. A motion vector associated with the current block based on the one or more retrieved motion vector predictors, and decoding the current block based on the determined motion vector. The reference motion vector candidate bank may be updated by inserting the motion vector associated with the current block into the reference motion vector candidate bank.