Image Decoding Device Adaptive Weighting Coefficients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geometric partitioning mode (GPM) techniques for image decoding have limited weighted averaging patterns, resulting in suboptimal encoding performance.
Innovation Solution
An image decoding device and method that selects unique weighting coefficients based on indirect control information to perform weighted averaging of predicted samples, allowing for variable division boundary widths and patterns, thereby improving encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If limited weighted averaging patterns are used in GPM, then device complexity is reduced, but encoding performance deteriorates
Solution Approach 1:
The patent applies dynamics by making the weighting coefficients adaptive rather than fixed. The weighting coefficients are determined based on block size and blurring conditions, allowing the system to dynamically adjust the blending behavior to match the specific characteristics of each video block, thereby improving encoding performance without significantly increasing complexity
Solution Approach 2:
The patent changes parameters by introducing multiple weighting coefficient patterns (first through fourth patterns) with different distribution characteristics. The system selects appropriate patterns based on block size and blurring conditions, effectively changing the weighting parameters to optimize encoding performance for different video content scenarios
2Device complexity
If fixed weighting coefficients are used, then encoding complexity is reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent employs parameter changes by defining multiple weighting coefficient patterns with different distribution characteristics (e.g., linear, quadratic, cubic). The appropriate pattern is selected based on block size and blurring conditions, allowing the system to adapt parameters to achieve higher prediction accuracy without significantly increasing encoding complexity
Solution Approach 2:
The patent applies local quality by using different weighting coefficient patterns for different regions and conditions. Specifically, different patterns are applied based on block size (e.g., 8×8, 16×16, 32×32) and blurring conditions, ensuring that each local region receives the most appropriate weighting treatment for its specific characteristics
3Measurement precision
If multiple weighting coefficient patterns are introduced, then prediction accuracy is improved, but encoding complexity increases
Solution Approach 1:
The patent manages complexity through parameter changes by organizing multiple weighting coefficient patterns into a systematic framework. The patterns are selected based on simple criteria (block size and blurring conditions), which allows the system to use multiple patterns without proportionally increasing encoding complexity, as the selection process itself is relatively straightforward
Solution Approach 2:
The patent reduces overall complexity by applying local quality - using different weighting patterns only where appropriate based on local block characteristics. Not all blocks require all patterns; the system selectively applies patterns based on block size and blurring conditions, thereby improving prediction accuracy only where needed while maintaining lower complexity for other blocks
Data Source
AI summary
In an image decoding device (200) according to the present invention, a circuit: decodes control information and a quantized value; obtains a decoded transform coefficient by performing inverse quantization on the decoded quantized value; obtains a decoded prediction residual by performing inverse transform on the decoded transform coefficient; generates a first predicted sample based on a decoded sample and the decoded control information; accumulates the decoded sample; generates a second predicted sample based on the accumulated decoded sample and the decoded control information; generates a third predicted sample by weighted averaging using weighting coefficients which are uniquely selected from among a plurality of weighting coefficients based on indirect control information for at least one of the first predicted sample or the second predicted sample; and obtains the decoded sample by adding the decoded prediction residual and the third predicted sample.