Video Coding Block Split Motion Information Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding, the geometric partitioning mode faces complexity in storing motion information for block units, as some block components are predicted based on one merge candidate while others are predicted based on both, leading to inefficient storage and prediction processes.
Innovation Solution
A method and device for decoding a bitstream that determines a block unit, splits it based on a split angle and distance, and selectively stores motion information for sub-blocks based on parameter values, allowing for efficient reconstruction of subsequent blocks by choosing between first and second motion information based on threshold comparisons.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If two merge candidates are stored for each sub-block predicted based on both merge candidates, then the motion prediction can be more accurate, but the complexity of stored motion candidates becomes too high
Solution Approach 1:
The patent applies local quality by differentiating the treatment of motion candidates based on sub-block location. Sub-blocks are categorized into first, second, and third types based on their positions relative to the split line. First and second type sub-blocks store only one merge candidate (reducing complexity), while third type sub-blocks store both merge candidates (maintaining accuracy where needed). This selective approach resolves the contradiction by optimizing the balance between prediction accuracy and storage complexity locally rather than uniformly across all sub-blocks.
2Reliability
If motion information is stored for all sub-blocks, then subsequent block reconstruction is more accurate, but the storage requirement and processing complexity increase
Solution Approach 1:
The patent extracts and stores motion information selectively rather than for all sub-blocks. By identifying that only third type sub-blocks (those straddling the split line) require both merge candidates for accurate reconstruction, the patent extracts and stores motion information only where necessary. This extraction principle reduces storage requirements and processing complexity while maintaining reconstruction accuracy for the critical sub-blocks that require it.
3Measurement precision
If both merge candidates are used for all block components, then prediction accuracy is improved, but the coding complexity and bitstream overhead increase
Solution Approach 1:
The patent segments the block components into different categories based on their relationship to the split line. By dividing sub-blocks into first, second, and third types, the patent enables different coding strategies for different segments. First and second type sub-blocks use simplified coding with one merge candidate, while third type sub-blocks use the more complex dual-merge-candidate approach. This segmentation resolves the contradiction by applying the appropriate level of complexity to each segment rather than uniformly across all block components.
Data Source
AI summary
A method of decoding a bitstream by an electronic device is provided. A block unit is determined from an image frame received from the bitstream. A split line of the block unit is determined based on a split angle and a split distance of the block unit determined based on the bitstream. First motion information and second motion information are received from a candidate list for reconstructing the block unit. A parameter value of a sub-block determined in the block unit is determined based on a location of the sub-block, the split angle and the split distance. A predefined one of the first and the second motion information is stored for the sub-block when the parameter value is greater than a first threshold and less than second threshold. A subsequent block is reconstructed based on the stored motion information when the subsequent block is reconstructed based on the sub-block.