Affine Motion Prediction Subblock Sizing for Video Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for a method and device that reduces the complexity and additional data generated by affine motion prediction in video signal processing, particularly in next-generation video content with high spatial resolution, frame rate, and scene representation, while minimizing performance degradation.
Innovation Solution
A video signal processing method and device that uses pre-defined subblock sizes for affine motion prediction, determining motion vectors for subblocks based on control points, and generating prediction samples, with specific size configurations for bi-prediction and uni-prediction modes to optimize block sizes and reduce computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If affine motion prediction is applied to represent various motions in video content, then motion representation accuracy is improved, but computational complexity and additional data generation increase
Solution Approach 1:
The current block is divided into multiple subblocks, with each subblock having its own motion vector derived from control point motion vectors. This segmentation allows the affine motion model to represent complex motions more accurately while managing computational complexity by processing smaller, independent subblocks rather than treating the entire block as a single unit.
Solution Approach 2:
Different subblocks within the current block can have different motion characteristics, and the affine motion prediction applies local motion vectors to each subblock based on its position and the control point motion vectors. This local quality approach enables accurate representation of varying motions across different regions of the block without requiring a single complex motion model for the entire block.
2Measurement precision
If affine motion prediction is applied to represent various motions in video content, then motion representation accuracy is improved, but additional data generation increases
Solution Approach 1:
The control point motion vectors serve multiple purposes: they are used to derive motion vectors for all subblocks within the current block, and they can also be used for motion compensation. This multi-functionality reduces the need for separate data structures and minimizes additional data generation while maintaining accurate motion representation across the entire block.
Solution Approach 2:
The patent changes the parameter of subblock size to pre-defined values (e.g., 4x4 or 8x8 pixels) based on the block type (bi-prediction or uni-prediction). This parameter change optimizes the balance between motion representation accuracy and data generation by using appropriate subblock sizes for different prediction modes, reducing unnecessary data for smaller blocks while maintaining accuracy for larger blocks.
3Device complexity
If pre-defined subblock sizes are used in affine motion prediction, then computational complexity is reduced, but adaptability to different block sizes decreases
Solution Approach 1:
The patent dynamically selects the subblock size based on the block type: 4x4 subblocks for uni-prediction blocks and 8x8 subblocks for bi-prediction blocks. This dynamic adaptation allows the system to use pre-defined subblock sizes (reducing complexity) while still adapting to different block characteristics (maintaining versatility). The subblock size is not fixed but adjusts according to the prediction mode being used.
Data Source
AI summary
A method for processing a video signal using an affine motion prediction is disclosed. The method includes checking that a current block is encoded by the affine motion prediction, obtaining motion vectors for a plurality of control points of the current block, determining a motion vector for each of a plurality of subblocks included in the current block based on the motion vectors for the plurality of control points, and performing a prediction for the current block from the motion vector for each of the plurality of subblocks, wherein each of the plurality of subblocks is configured to have a pre-defined width and a pre-defined height.