Multi-Predictor Combination for Video Block Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The High-Efficiency Video Coding (HEVC) standard faces challenges in coding efficiency due to parsing errors and redundancy in motion information, particularly in Advanced Motion Vector Prediction (AMVP), Skip, and Merge modes, which can lead to parsing errors and propagate through subsequent pictures.
Innovation Solution
The proposed solution involves generating a final predictor for a current block by combining multiple predictors using settings such as combined-weighting, applied-portion, motion information, and precision settings, allowing for improved prediction accuracy through multi-hypothesis mode, which includes inter prediction modes like Skip, Merge, AMVP, affine Merge, and affine Inter modes, and intra prediction modes, thereby enhancing prediction accuracy and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple predictors are combined for block prediction, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the prediction process by dividing the current block into multiple sub-blocks and generating different predictors for each sub-block based on different hypotheses (intra prediction, inter prediction, merge mode, AMVP mode). This segmentation allows combining multiple prediction results without processing the entire block uniformly, improving accuracy while managing complexity through localized processing.
Solution Approach 2:
The patent merges multiple predictors generated from different hypotheses (intra prediction predictor, inter prediction predictor, merge mode predictor, AMVP mode predictor) using weighted combination. The final prediction for each sub-block is obtained by combining these predictors with different weights, thereby improving prediction accuracy by leveraging multiple sources of prediction information.
2Productivity
If multiple predictors are combined for block prediction, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent introduces dynamic weight adjustment based on block characteristics. The weights assigned to different predictors are not fixed but are determined dynamically according to the block size, block position, and prediction mode. This dynamic approach allows the system to adapt to different coding scenarios, improving coding efficiency while avoiding unnecessary complexity in uniform scenarios.
Solution Approach 2:
Different sub-blocks within the current block may use different combinations of predictors and different weightings based on their local characteristics. This local quality approach allows the system to apply more complex prediction strategies only where needed, improving overall coding efficiency while limiting the increase in device complexity to specific regions rather than the entire block.
3Measurement precision
If motion information is extensively used for prediction, then prediction accuracy is improved, but parsing errors and redundancy increase
Solution Approach 1:
The patent changes the parameter representation by using multiple hypotheses with different motion information sources (spatial neighbors, temporal references, combined modes) instead of relying on a single motion vector. This parameter diversification allows the system to select or combine predictors based on which hypothesis provides the most reliable prediction, reducing the impact of parsing errors in any single motion information source.
Data Source
AI summary
Video processing methods and apparatuses for coding a current block generate a final predictor by combining multiple predictors for the current block. A first predictor and a second predictor for the current block are generated by applying one or a combination of settings to the first, second, or both predictors. One or both the first and second predictors are generated from motion compensation. The final predictor is derived from the first and second predictors, and the current block is encoded or decoded according to the final predictor. The settings include supported-mode setting, combined-weight setting, applied-portion setting, motion information setting, precision setting, or a combination of the above settings.