Variable Weight Intra Prediction for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video encoding and decoding methods, such as those in H.265/HEVC, rely on fixed weights for multi-reference intra prediction, which limits the efficiency of predicting video blocks, especially with larger block sizes and increased angular prediction directions, leading to suboptimal compression efficiency.
Innovation Solution
The method allows the encoder to select and transmit different sets of weights for each block, enabling variable weight application to multiple reference arrays for intra prediction, improving prediction accuracy and compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed weights are used for multi-reference intra prediction, then device complexity is reduced, but compression efficiency deteriorates
Solution Approach 1:
The patent transforms the static fixed-weight approach into a dynamic system where weights are selected adaptively based on block characteristics. The encoder evaluates multiple candidate weight sets and selects the optimal one for each block, allowing the prediction system to adapt to varying content types and block sizes, thereby improving compression efficiency while maintaining manageable complexity through structured selection processes.
Solution Approach 2:
The patent changes the weight parameters from fixed predetermined values to variable parameters that can be selected from multiple candidate sets. By introducing multiple weight sets (e.g., different angular prediction weights, different block size-specific weights) and selecting appropriate parameters based on block characteristics, the system achieves better compression efficiency without requiring complete redesign of the prediction framework.
2Productivity
If variable weights are selected and transmitted for each block, then compression efficiency is improved, but bitrate increases
Solution Approach 1:
The patent applies partial action by selectively transmitting weight information only when it provides meaningful improvement. Instead of transmitting full weight sets for every block, the system uses conditional transmission based on block characteristics, content complexity, and potential gain thresholds. This approach achieves compression efficiency improvements while minimizing the additional bitrate overhead by applying variable weights only where beneficial.
Solution Approach 2:
The patent performs preliminary evaluation at the encoder to determine which weight sets will provide actual benefit before transmitting them. The encoder assesses block characteristics, predicts potential compression gains, and pre-selects which weight information to transmit. This preliminary action prevents unnecessary bitrate consumption by avoiding transmission of weight data that would not contribute meaningfully to compression efficiency.
3Measurement precision
If multiple reference arrays are used with variable weights, then prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the weight selection process into manageable components: block characteristic analysis, candidate weight set evaluation, and optimal weight selection. By dividing the complex prediction process into these distinct segments, the system can handle multiple reference arrays with variable weights in a structured manner, improving prediction accuracy while keeping implementation complexity manageable through modular processing steps.
Data Source
AI summary
A method of performing intra prediction for encoding or decoding uses multiple layers of reference samples. The layers are formed into reference arrays that are used by a function, such as a weighted combination, to form a final prediction. The weights can be variable, chosen from among a number of number of sets of weights. The prediction is used in encoding or decoding a block of video data. The weights can be determined in a number of ways, and for a given prediction mode, the same weights, or different weights can be used for all pixels in a target block. If the weights are varied, they can depend on the distance of the target pixel from reference arrays. An index can be sent indicating which set of weights is to be used.