IBC Block Vector Precision Selection Using Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in block vector resolution during intra block copy (IBC) mode, particularly in advanced video coding standards like VVC, where adaptive motion vector resolution is not optimized for template matching, leading to suboptimal compression performance.
Innovation Solution
Implement template-matching based adaptive block vector resolution (ABVR) in IBC mode, which reorders block vector precisions based on template matching differences to select the most accurate precision for block vector reconstruction, enhancing compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If adaptive motion vector resolution is used in IBC mode, then compression efficiency is improved, but the complexity of block vector resolution selection increases
Solution Approach 1:
The patent changes the parameter of block vector precision from a fixed value to an adaptive parameter that can take multiple precision levels (e.g., 1/4-pel, 1/2-pel, 1-pel). The encoder selects different precision levels based on template matching results, allowing the system to adaptively adjust the resolution parameter to balance compression efficiency and computational complexity for different block types and content characteristics.
Solution Approach 2:
The patent introduces dynamic selection of block vector precision levels based on template matching performance. Instead of using a static resolution for all blocks, the system dynamically adjusts the precision requirement for each block vector depending on the matching accuracy needed, making the resolution selection process flexible and adaptive to different coding scenarios.
2Measurement precision
If multiple block vector precisions are used, then reconstruction accuracy is improved, but the signaling overhead increases
Solution Approach 1:
The patent performs template matching and evaluates different precision levels in advance during the encoding process. By pre-assessing which precision level will provide the best reconstruction accuracy for each block, the system can make informed decisions about which precision to use, avoiding the need to signal all possible precision options and reducing the overall signaling overhead.
Solution Approach 2:
The patent applies different block vector precision levels to different blocks based on their specific requirements. Instead of using a uniform high precision for all blocks (which would increase signaling overhead), the system locally adapts the precision level to match the actual reconstruction needs of each block, thereby maintaining accuracy where needed while reducing overhead where lower precision suffices.
Data Source
AI summary
Coding information is received and indicates that an intra block copy (IBC) mode is applied to a current block. Based on the IBC mode being applied to the current block, first adaptive block vector resolution (ABVR) information included in the coding information is obtained from the received video bitstream. The first ABVR information is determined to indicate that a plurality of BV precisions is associated with a block vector (BV) of the IBC mode. The plurality of BV precisions associated with the BV of the current block is reordered based on template matching (TM) differences between a template region of the current block and each of a plurality of template regions of reference blocks. A particular BV precision is selected from the plurality of reordered BV precisions by the TM differences. The current block is reconstructed based at least on the selected particular BV precision.