Affine Motion Compensation Range Limits for Video Encoders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in motion estimation and compensation, particularly in affine motion compensation prediction, leading to high memory bandwidth requirements and variations in control point motion vectors.
Innovation Solution
Implementing a system that limits the range of motion estimation and compensation in affine motion compensation prediction, reducing variations between control point motion vectors and optimizing memory bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion estimation and compensation are performed without range limits, then prediction accuracy is improved, but memory bandwidth requirements increase
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different regions in the reference picture. A motion estimation range is defined that includes only the region relevant to the current block, while other regions are excluded. This allows high-precision motion estimation for the current block without requiring access to the entire reference picture, thereby reducing memory bandwidth requirements.
2Loss of energy
If the motion estimation range is limited, then memory bandwidth requirements are reduced, but the variation in control point motion vectors increases
Solution Approach 1:
The patent applies preliminary action by pre-defining the motion estimation range based on the current block's position and size before performing motion estimation. The range is determined in advance using the block's top-left and bottom-right coordinates, ensuring that the motion estimation is performed efficiently within a predetermined boundary. This preliminary range definition helps control the variation in control point motion vectors while maintaining low memory bandwidth requirements.
3Measurement precision
If affine motion compensation is applied to all blocks, then prediction accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent applies local quality by selectively applying affine motion compensation based on the current block's characteristics. The encoder determines whether to use affine motion compensation for each block based on factors such as block size and motion characteristics. This allows high-precision prediction for blocks that benefit from affine compensation while avoiding the processing complexity for blocks where simpler motion models are sufficient.
Data Source
AI summary
An encoder includes: circuitry; and memory, in which using the memory, the circuitry, in affine motion compensation prediction in inter prediction for a current block, places a limit on a range within which motion estimation or motion compensation is performed, and performs the motion compensation for the current block.