Adaptive Motion Vector Resolution for Video Block Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video signal processing methods lack efficiency in coding and signaling of motion information, particularly in motion vector resolution, which affects overall video compression and transmission quality.
Innovation Solution
A method and apparatus for video signal processing that configures a Motion Vector Prediction (MVP) candidate list using two methods, adjusts motion vector differential values based on resolution sets specific to each method, and reconstructs blocks using modified motion vectors, enhancing coding and signaling efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a single resolution is used for motion vector differential values, then device complexity is reduced, but coding efficiency deteriorates
Solution Approach 1:
The patent applies dynamics by making the resolution of motion vector differential values adaptive rather than fixed. The resolution is dynamically selected from multiple available resolutions based on the characteristics of the current block and the configured MVP candidate list method, allowing the system to optimize coding efficiency for different block types while maintaining manageable device complexity through standardized adaptation rules.
Solution Approach 2:
The patent changes the parameter of motion vector differential value resolution from a fixed single value to a selectable set of multiple resolutions. By allowing the resolution parameter to vary according to block characteristics and MVP configuration methods, the system achieves improved coding efficiency without requiring completely new processing architectures, thus avoiding excessive device complexity.
2Loss of information
If multiple resolution sets are used for different MVP configuration methods, then signaling efficiency is improved, but device complexity increases
Solution Approach 1:
The patent segments the resolution selection into two distinct resolution sets: a first resolution set for when the MVP candidate list is configured through a first method, and a second resolution set for when configured through a second method. This segmentation allows each resolution set to be optimized for its specific MVP configuration method, improving signaling efficiency while keeping the complexity manageable through clear separation of cases.
Solution Approach 2:
The patent introduces parameter changes by defining different resolution sets based on the MVP configuration method. The encoder and decoder can adaptively select appropriate resolutions from the appropriate set based on the configured MVP method, achieving better signaling efficiency without requiring complex real-time resolution negotiation, as the resolution set is determined by the already-configured MVP method.
3Productivity
If motion vector resolution is optimized for each block type, then coding efficiency increases, but processing complexity increases
Solution Approach 1:
The patent applies local quality by allowing different motion vector differential value resolutions to be applied to different current blocks based on their specific characteristics and the MVP configuration method used for each block. This enables coding efficiency optimization at the local block level while maintaining overall system manageability through consistent adaptation rules that are applied uniformly across all blocks.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
An apparatus for decoding a video signal, the apparatus comprising a processor, wherein the processor is configured to: obtain a motion vector difference value related to a current block, obtain a resolution related to the motion vector difference value, obtain a modified motion vector difference value by modifying the motion vector difference value based on the resolution, obtain a motion vector of the current block based on the modified motion vector difference value, and reconstruct the current block based on the motion vector. The apparatus is characterised by: wherein the resolution is one of a first plurality of resolutions included in a first resolution set or one of a second plurality of resolutions included in a second resolution set, i) when whether the resolution is selected from the first plurality of resolutions or from the second plurality of resolutions is determined based on whether a prediction mode of the current block is an affine mode: the first plurality of resolutions and the second plurality of resolutions both include a resolution in unit of 1 sample, the first plurality of resolutions does not include a largest resolution among the second plurality of resolutions, ii) when whether the resolution is selected from the first plurality of resolutions or from the second plurality of resolutions is determined by using a picture order count, POC, of a current picture including the current block and a POC of a reference picture of the current picture, the first plurality of resolutions and the second plurality of resolutions both include a resolution in unit of 1 sample, the first plurality of resolutions includes a resolution in unit of 1/2 sample, and the second plurality of resolutions does not include the resolution in unit of 1/2 sample.