Intra Block Copy Prediction Using Non-Adjacent Neighboring Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards inefficiently utilize block vectors of non-adjacent neighboring blocks for prediction, particularly in constructing IBC merge lists and AMVP lists, and impose unreasonable constraints on block sizes and search ranges, leading to suboptimal coding efficiency.
Innovation Solution
The proposed solution involves using block vectors of non-adjacent neighboring blocks for prediction, adjusting their positions and classifications based on the current block's dimensions and positions, and modifying the search range constraints to enhance prediction accuracy and efficiency, particularly by allowing non-adjacent blocks from different CTU rows and relaxing strict constraints on block sizes and search ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If block vectors of non-adjacent neighboring blocks are used for prediction in IBC merge lists and AMVP lists, then prediction accuracy and coding efficiency are improved, but the complexity of constructing these lists increases
Solution Approach 1:
The patent segments the process of constructing IBC merge lists and AMVP lists by introducing specific rules for selecting non-adjacent neighboring blocks based on CTU row relationships. This segmentation allows systematic inclusion of relevant blocks while excluding unnecessary ones, managing complexity through structured classification.
Solution Approach 2:
The patent applies local quality by differentiating the treatment of non-adjacent neighboring blocks based on their specific spatial relationships with the current block. Different selection criteria are applied to blocks in different CTU rows, allowing optimization for each local case while maintaining overall system efficiency.
2Productivity
If strict constraints on block sizes and search ranges are imposed, then the processing complexity is reduced, but the prediction accuracy and coding performance deteriorate
Solution Approach 1:
The patent changes the parameters of block size constraints and search range limitations to be more flexible. By modifying these parameters to allow larger blocks and extended search ranges, the system achieves better prediction accuracy without proportionally increasing processing complexity.
Solution Approach 2:
The patent introduces dynamic adjustment of search range constraints based on the specific characteristics of non-adjacent neighboring blocks. The search range is not fixed but adapts to the actual spatial relationship and content characteristics, allowing optimization for each case while maintaining manageable processing complexity.
3Adaptability or versatility
If non-adjacent neighboring blocks from different CTU rows are included in predictions, then the versatility of prediction sources is improved, but the reliability of prediction decreases due to increased search space
Solution Approach 1:
The patent segments the search space for non-adjacent neighboring blocks by CTU row, allowing systematic exploration of different regions. This segmentation maintains versatility by considering multiple CTU rows while managing reliability through structured search patterns that prioritize more promising regions.
Solution Approach 2:
The patent applies partial action by selectively including non-adjacent neighboring blocks from different CTU rows based on their potential usefulness. Not all possible blocks are considered, but rather a strategically selected subset that balances versatility with prediction reliability, avoiding the excessive action of examining every possible block.
Data Source
AI summary
Methods, systems, and devices for intra block copy (IBC) using non-adjacent neighboring blocks are described. An example method of video processing includes determining, for a conversion between a video including a current video block and a bitstream of the video, whether to use a block vector of a non-adjacent neighboring block for a prediction of a block vector of the current video block based on coding tree unit (CTU) rows that include the non-adjacent neighboring block and the current video block, and performing, based on the determining, the conversion.