Subblock Motion Compensation Filters for Bi-Prediction Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards like ITU-T H.264, ITU-T H.265, and ITU-T H.266 face challenges in efficiently performing motion compensation, particularly in cases where subblocks are predicted using uni-prediction or bi-prediction, due to the lack of optimized interpolation filters that account for subblock dimensions.
Innovation Solution
The use of cos-windowed sinc function-based motion compensation interpolation filters, which are selectively applied based on whether a subblock is predicted using uni-prediction or bi-prediction, and whether specific conditions based on subblock width and height are met, providing distinct window and scale configurations for these filters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single motion compensation interpolation filter is used for all subblocks, then the device complexity is reduced, but the manufacturing precision of motion compensation deteriorates
Solution Approach 1:
The patent applies different interpolation filters (first filter for uni-prediction, second filter for bi-prediction) to different prediction modes and subblock configurations. This local differentiation optimizes motion compensation accuracy for each specific case without requiring complex global filter selection mechanisms.
Solution Approach 2:
The patent changes filter parameters (window size, scale factors) based on subblock dimensions and prediction mode. By adapting filter parameters to local conditions rather than using fixed parameters, the system achieves higher precision without proportionally increasing complexity.
2Manufacturing precision
If distinct interpolation filters are selected for uni-prediction and bi-prediction based on subblock dimensions, then the manufacturing precision of motion compensation is improved, but the device complexity increases
Solution Approach 1:
The patent implements local quality by selecting different interpolation filters based on the specific prediction mode (uni-prediction or bi-prediction) and subblock dimensions. Each local configuration receives an optimized filter, improving overall accuracy without requiring a complex centralized decision-making system.
Solution Approach 2:
The system changes filter parameters dynamically based on detectable properties (prediction mode, subblock size) rather than maintaining fixed parameters. This allows precision improvement through adaptive parameter selection while keeping the complexity manageable through rule-based parameter adjustment.
3Productivity
If motion compensation uses optimized interpolation filters for different prediction methods, then the productivity of video coding is improved, but the device complexity increases
Solution Approach 1:
The patent optimizes video coding productivity by changing interpolation filter parameters based on prediction mode and subblock dimensions. This adaptive approach improves coding efficiency by reducing residual errors, while the parameter changes follow systematic rules that prevent exponential complexity growth.
Data Source
AI summary
A device may be configured to determine motion compensation interpolation filter. The device may be configured to select a motion compensation interpolation filter based on whether a block is predicted using uni-prediction or bi-prediction and whether a condition based on a subblock width and a subblock height is satisfied. The motion compensation interpolation filter may be defined according to cos-windowed sinc function.