Image Motion Compensation with Long-Term Reference Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding technologies fail to consider the particularity of long-term reference images during motion vector derivation and bidirectional motion prediction, leading to reduced search efficiency and encoding efficiency due to unclear temporal distances and artificial construction of reference images.
Innovation Solution
A method and device for determining a long-term reference image based on a frame identifier, updating it using a short-term reference image, and scaling candidate motion vectors appropriately to improve prediction accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion vector derivation is performed without considering long-term reference image characteristics, then encoding process is simpler, but search efficiency and encoding efficiency are reduced
Solution Approach 1:
The patent applies local quality by differentiating the treatment of long-term reference images from short-term reference images. Specifically, it identifies blocks in long-term reference images that have the same prediction direction and motion vector as the current block, and applies motion compensation only to those matching blocks. This selective approach improves encoding efficiency by avoiding unnecessary motion searches while maintaining accuracy where needed.
Solution Approach 2:
The patent implements preliminary action by pre-identifying and marking blocks in long-term reference images that share the same prediction direction and motion vector characteristics with the current block. This pre-processing allows the decoder to directly reuse motion vectors without performing full motion searches, thereby improving search efficiency and reducing computational complexity.
2Manufacturing precision
If temporal distance is not considered for long-term reference images, then reference image construction is simpler, but prediction accuracy deteriorates
Solution Approach 1:
The patent applies parameter changes by introducing a frame distance parameter to differentiate between long-term and short-term reference images. It uses this parameter to determine whether to perform motion compensation on reference blocks. By changing the handling parameter based on frame distance, the patent improves prediction accuracy while managing the complexity of reference image construction.
3Productivity
If motion search is performed on all reference images, then prediction accuracy is maintained, but search efficiency is reduced
Solution Approach 1:
The patent extracts and identifies specific blocks in long-term reference images that have identical prediction directions and motion vectors to the current block. By extracting only these relevant blocks for motion compensation rather than performing searches on all reference image blocks, the patent significantly improves search efficiency while maintaining prediction accuracy for the extracted blocks.
Data Source
AI summary
A video processing method includes obtaining a first initial motion vector (MV) and a second initial MV. The first initial MV points to a first reference image, and the second initial MV points to a second reference image. The method further includes, in response to the first reference image and the second reference image being both short-term reference images, calculating a motion offset of a current image block based on gradients at sampling points pointed to by the first initial MV and the second initial MV, and calculating a predicted image block of the current image block based on the motion offset of the current image block.