Multi-Hypothesis Video Prediction for Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-Efficiency Video Coding (HEVC) faces limitations in coding efficiency due to its reliance on single hypothesis prediction modes for inter-prediction, which can lead to suboptimal performance in certain block sizes and configurations, particularly in handling motion information for pixel blocks.
Innovation Solution
The implementation of a Multi-Hypothesis mode that generates a second prediction hypothesis using either intra-prediction or additional inter-prediction modes, allowing for combined predictions that improve reconstruction accuracy by considering multiple candidate lists based on block size, shape, and motion candidate directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If single hypothesis prediction mode is used for inter-prediction in HEVC, then device complexity is reduced, but coding efficiency deteriorates
Solution Approach 1:
The patent implements a dynamic prediction system that adaptively switches between single hypothesis and multi hypothesis modes based on block characteristics. The decoder determines whether to generate a second prediction based on block size, shape, and motion characteristics, making the system flexible and adaptive rather than static
Solution Approach 2:
The patent changes the parameter of prediction hypothesis count from fixed (single) to variable (single or multiple). By modifying the number of prediction hypotheses generated based on block properties, the system optimizes coding efficiency while controlling complexity through parameter adaptation
2Productivity
If multiple hypothesis prediction modes are implemented, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies multi-hypothesis prediction selectively to specific blocks based on their local characteristics such as size, shape, and motion properties. Not all blocks receive multiple hypotheses - only those that benefit from it, making the complexity localized rather than universal
Solution Approach 2:
The patent implements partial multi-hypothesis action by generating a second prediction only for certain blocks where it provides benefit. The decoder uses criteria such as block size thresholds and motion candidate directions to determine when to apply the additional hypothesis, avoiding unnecessary complexity for blocks where single hypothesis suffices
3Measurement precision
If combined prediction from multiple hypotheses is used, then prediction accuracy is improved, but computational cost increases
Solution Approach 1:
The patent applies combined prediction partially - only for blocks where the second hypothesis is generated. The computational cost of combining predictions is incurred only when beneficial, based on block characteristics, rather than for all blocks uniformly
Data Source
AI summary
A video decoder receives to-be-decoded data from a bitstream for a block of pixels to be decoded as a current block of a current picture of the video. The video decoder generates a first prediction of the current block based on a first prediction mode that is selected from a first candidate list. The video decoder generates a second prediction of the current block based on a second prediction mode that is selected from a second candidate list. The video decoder generates a combined prediction for the current block based on the first prediction and the second prediction. The video decoder reconstructs the current block by using the combined prediction.