Intra Block Copy Vectors With Adaptive Subpixel Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods for screen content, which lack subpixel precision in expressing block vectors, result in low coding efficiency due to the inherent differences between natural and screen content characteristics.
Innovation Solution
Adaptive signaling of block vector spatial resolution and block vector difference sign to improve coding efficiency by allowing subpixel precision in block vector representation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional encoding methods are used for screen content, then the encoding process is simple, but coding efficiency is low due to lack of subpixel precision in block vector representation
Solution Approach 1:
The patent implements dynamic adaptive resolution by allowing the block vector precision to be adjusted according to the content type and coding requirements. The system can switch between different precision levels (e.g., pixel-level and subpixel-level) based on the characteristics of the screen content, making the encoding process adaptable rather than static.
Solution Approach 2:
The patent changes the precision parameter of block vectors from fixed pixel-level to variable subpixel-level precision. By introducing fractional pixel positions (e.g., 0.5-pel, 0.25-pel), the system enhances the measurement precision of block vector representation while maintaining encoding efficiency through selective application.
2Productivity
If subpixel precision is implemented in block vector representation, then coding efficiency improves, but the complexity of signaling and processing increases
Solution Approach 1:
The patent applies subpixel precision locally only where needed in the screen content, rather than uniformly across all blocks. By identifying regions that benefit from higher precision (e.g., areas with fine details or gradients) and applying subpixel block vectors selectively, the system improves coding efficiency in those specific areas while avoiding unnecessary complexity in other regions.
Solution Approach 2:
The patent implements partial subpixel precision by using fractional pixel positions (e.g., 0.5-pel, 0.25-pel) only when beneficial for coding efficiency. The system does not apply maximum precision universally but rather uses the appropriate level of precision partiality based on content characteristics, balancing improved productivity with controlled complexity.
3Adaptability or versatility
If block vector difference is signaled with fixed spatial resolution, then the signaling process is simple, but the ability to represent subpixel positions is limited
Solution Approach 1:
The patent implements dynamic resolution signaling where the spatial resolution of block vector differences adapts based on the coding context. The system can switch between different resolution levels (e.g., pixel-level and subpixel-level) depending on the content characteristics and coding requirements, making the resolution signaling flexible and adaptive rather than fixed.
Solution Approach 2:
The patent changes the spatial resolution parameter of block vector differences from fixed to variable. By introducing subpixel-level precision (e.g., 0.5-pel, 0.25-pel) as an additional resolution level, the system enhances adaptability while managing complexity through selective application based on content needs.
Data Source
AI summary
A video decoding apparatus for adaptively generating a block vector of a current block in an intra block copy (IBC) mode includes at least one processor. The at least one processor is configured to decode a block vector predictor index and an absolute value of a block vector difference from a bitstream; generate a block vector predictor candidate list of the current block and generating a block vector predictor from the block vector predictor candidate list using the block vector predictor index; derive or decode a sign of the block vector difference using the block vector predictor; generate the block vector difference by combining the absolute value and the sign of the block vector difference; and generate the block vector by combining the block vector predictor and the block vector difference.