Pyramidal Motion Estimation Hints for Multi-Object Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding techniques using pyramidal motion estimation struggle with accurately estimating the motion of multiple objects in high-resolution media frames, leading to poor visual quality due to shared motion vector hints and inadequate estimation of object motion direction and speed changes.
Innovation Solution
The method involves dividing media frames into smaller grid blocks and generating multiple motion vector hints for each block, with separate hints for forward and backward prediction, improving motion estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pyramidal motion estimation is used to reduce computational complexity, then encoding speed is improved, but motion estimation accuracy deteriorates due to shared motion vector hints
Solution Approach 1:
The patent divides each block into multiple sub-blocks and generates separate motion vector hints for each sub-block instead of sharing a single motion vector hint. This segmentation allows each sub-block to have its own motion estimation, improving accuracy while maintaining the computational efficiency of pyramidal motion estimation by still using the hierarchical phase approach.
Solution Approach 2:
The patent applies different motion vector hints to different sub-blocks within a block, allowing each sub-block to have locally optimized motion estimation. This local quality approach ensures that regions with different motion characteristics (e.g., multiple objects moving in different directions) are handled appropriately, improving overall motion estimation accuracy.
2Device complexity
If a single motion vector hint is shared for forward and backward motion estimation, then device complexity is reduced, but motion estimation accuracy deteriorates in scenes with rapidly changing objects
Solution Approach 1:
The patent generates separate motion vector hints for forward and backward motion estimation for each sub-block, rather than sharing a single hint. This segmentation of hint data allows independent optimization for forward and backward prediction, improving accuracy in scenes with rapidly changing objects while maintaining reasonable data structure complexity.
Solution Approach 2:
The patent dynamically generates different motion vector hints based on the motion characteristics of each sub-block and the prediction direction (forward or backward). This dynamic approach allows the system to adapt to rapidly changing object motion by providing direction-specific and region-specific motion estimates.
Data Source
AI summary
Various embodiments include techniques for performing hardware-assisted motion estimation when encoding blocks of a media frame. A video encoder generates motion vector hints for multiple phases of motion estimation at increasingly finer resolutions until the final motion estimation phase at the full resolution of the media frame. The video encoder divides the media frame into a grid of grid blocks, where the grid blocks are smaller than the standard encoding block size of the media frame. At each phase, the video encoder generates one motion vector hint for each grid block at the selected resolution for that phase. For each motion vector hint, the video encoder generates forward prediction hint data based on a reference media frame previous to the current media frame. Further, for each motion vector hint, the video encoder generates backward prediction hint data based on a reference media frame that follows the current media frame.