Merge Candidate List Generation for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In High Efficiency Video Coding (HEVC), the generation of merge candidate lists for inter-picture prediction is limited, leading to suboptimal bit rate efficiency and encoding performance, particularly in scenarios requiring advanced motion vector prediction modes.
Innovation Solution
The proposed method involves determining specific candidate blocks, including a first candidate block above the current block's top edge and a second candidate block to the left of the current block's left edge, to generate a more comprehensive merge candidate list, enhancing the encoding and decoding of video sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If the merge candidate list is limited to traditional spatial and temporal candidates, then the encoding complexity is reduced, but the bit rate efficiency and prediction performance deteriorate
Solution Approach 1:
The patent introduces middle candidates positioned at the center of current block boundaries, adding a new spatial dimension to candidate selection. This extends beyond traditional corner-based spatial candidates and temporal candidates, creating a more comprehensive three-dimensional candidate space that improves prediction accuracy without excessively increasing complexity
2Measurement precision
If additional middle candidates are added to the merge candidate list, then the motion vector prediction accuracy is improved, but the encoding complexity increases
Solution Approach 1:
The patent applies local quality by positioning middle candidates specifically at the center locations of current block boundaries where motion information is most relevant. This targeted approach ensures that additional candidates are placed where they provide maximum predictive value, improving local prediction accuracy without uniformly increasing complexity across all block types
3Productivity
If a comprehensive merge candidate list is generated with middle candidates, then the video compression efficiency is enhanced, but the processing time increases
Solution Approach 1:
The patent performs preliminary action by pre-positioning middle candidates at standardized center locations of block boundaries before the actual merge mode selection process. This advance preparation of candidate positions allows for faster evaluation during encoding, as the spatial relationships and candidate locations are predetermined, reducing real-time processing overhead
Data Source
AI summary
A method and apparatus for generating a merge candidate list for encoding or decoding of a video sequence includes determining a first candidate block that includes a vertical position above a top edge of a current block, and that includes a horizontal position adjacent to a horizontal center of the current block. A second candidate block is determined that includes a horizontal position that is located to a left side of a left edge of the current block and a vertical position adjacent to a vertical center of the current block. The merge candidate list is generated using the first candidate block and the second candidate block to permit the encoding or the decoding of the video sequence.