Adaptive Motion Vector Resolution Using Template Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently resolving motion vectors, particularly in adapting to varying resolutions and prediction modes, which affects compression efficiency and bitstream size.
Innovation Solution
The implementation of adaptive motion vector resolution (AMVR) mode using template matching (TM) to determine the best motion vector resolution for each block, by calculating TM costs and reordering available resolutions for optimal prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed motion vector resolution is used for all blocks, then the device complexity is reduced and processing is simplified, but the prediction accuracy decreases and compression efficiency is compromised
Solution Approach 1:
The patent implements adaptive motion vector resolution (AMVR) that dynamically selects the motion vector resolution for each block based on template matching results. The system evaluates template matching costs at different resolutions and adapts the resolution selection to the specific characteristics of each block, allowing the motion vector processing to transition from a static fixed resolution approach to a dynamic adaptive approach that optimizes prediction accuracy while managing complexity.
Solution Approach 2:
The patent applies different motion vector resolutions to different blocks based on their local characteristics. By performing template matching on a block-by-block basis and selecting the appropriate resolution for each individual block, the system ensures that each local region is processed with the optimal resolution, rather than applying a uniform resolution across the entire picture.
2Manufacturing precision
If template matching is performed on all available motion vector resolutions, then the prediction accuracy is improved, but the computational cost and processing time increase
Solution Approach 1:
The patent performs template matching evaluations in advance for different motion vector resolutions to determine the optimal resolution for each block before actual motion compensation. By pre-evaluating the template matching costs at various resolutions and storing these results, the system avoids redundant computations during the actual decoding process, reducing overall processing time while maintaining prediction accuracy.
Solution Approach 2:
The patent divides the template matching process into separate evaluations for different motion vector resolutions. Instead of performing a single comprehensive search, the system segments the evaluation into distinct template matching operations at 1/4-pel, 1/2-pel, 1-pel, and 4-pel resolutions, allowing for more efficient processing and selection of the optimal resolution for each block.
3Manufacturing precision
If higher motion vector resolution is used for all blocks, then the prediction accuracy is improved, but the bitstream size increases due to additional data requirements
Solution Approach 1:
The patent changes the motion vector resolution parameter adaptively for each block based on template matching results. Instead of using a fixed high resolution for all blocks, the system adjusts the resolution parameter according to the specific characteristics of each block, using higher resolutions only where necessary to maintain prediction accuracy while reducing the bitstream size for blocks that can be adequately represented at lower resolutions.
4Productivity
If adaptive motion vector resolution is implemented, then the compression efficiency is improved, but the device complexity increases due to additional processing requirements
Solution Approach 1:
The patent implements a self-service mechanism where the template matching process automatically determines the optimal motion vector resolution for each block without requiring external intervention or complex control logic. The system uses the template matching costs themselves to drive the resolution selection, allowing the process to self-regulate and adapt to different block characteristics while maintaining compression efficiency.
Data Source
AI summary
Aspects of the disclosure provide methods and apparatuses for video coding. A method for video decoding includes receiving a coded video bitstream that includes prediction information of a current block in a current picture. The prediction information indicates that an adaptive motion vector prediction (AMVP) mode and an adaptive motion vector resolution (AMVR) mode are applied to the current block. A set of motion vector (MV) resolutions in the AMVR mode is available for the current block. The method includes determining respective template matching (TM) costs corresponding to each of a plurality of MV resolutions in the set of MV resolutions based on the plurality of MV resolutions and motion information of the current block, reordering the plurality of MV resolutions based on the determined TM costs, selecting an MV resolution from the reordered plurality of MV resolutions, and reconstructing the current block based on the selected MV resolution.