Motion Vector Refinement for Inter Prediction in Image Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-resolution and high-quality images require efficient compression techniques to reduce transmission and storage costs, as conventional methods result in increased data amounts, necessitating improved image coding efficiency and motion vector accuracy while minimizing side information.
Innovation Solution
The method involves refining motion vectors using adaptive motion vector resolution, deriving accurate motion vectors by processing integer and fractional sample units, and generating predicted samples based on neighboring blocks, thereby enhancing inter prediction efficiency and reducing data allocation to motion vector differences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional image compression techniques are used for high-resolution images, then transmission and storage costs increase, but image quality and resolution are maintained
Solution Approach 1:
The patent applies adaptive motion vector resolution by dynamically changing the precision parameter of motion vectors based on block characteristics. For blocks with small motion changes, lower precision (integer sample units) is used, while for blocks with large motion changes, higher precision (fractional sample units) is applied. This parameter adaptation reduces overall data amount without compromising image quality.
Solution Approach 2:
The patent implements local quality control by applying different motion vector precision levels to different blocks within the image based on their specific characteristics. Each block is evaluated independently, and the appropriate motion vector resolution is applied locally, ensuring that only necessary areas consume high precision while other areas use compressed representations.
2Measurement precision
If motion vector precision is increased to improve inter prediction accuracy, then coding efficiency improves, but side information requirements increase
Solution Approach 1:
The patent dynamically adjusts the precision parameter of motion vectors based on block characteristics such as motion complexity and prediction accuracy requirements. By changing the parameter (integer vs. fractional sample units) adaptively, the system achieves high measurement precision only where necessary, reducing overall side information requirements.
Solution Approach 2:
The patent applies high-precision motion vectors only to specific blocks that require it, rather than uniformly applying high precision to all blocks. This partial action approach ensures that measurement precision is improved where needed while avoiding the excessive side information overhead that would result from universal high-precision application.
3Measurement precision
If uniform high-precision motion vectors are applied to all blocks, then prediction accuracy improves, but coding complexity and data requirements increase
Solution Approach 1:
The patent applies different levels of motion vector precision to different blocks based on their local characteristics. Blocks with simple motion patterns use lower precision (reducing complexity), while blocks with complex motion use higher precision (maintaining accuracy). This local differentiation resolves the contradiction between uniform accuracy and variable complexity.
Solution Approach 2:
The patent applies high-precision motion vectors only partially to blocks that require it, rather than excessively applying high precision uniformly. This selective approach maintains prediction accuracy where needed while significantly reducing coding complexity and data requirements for blocks where high precision is not necessary.
Data Source
AI summary
An inter prediction method performed by a decoding apparatus includes: receiving information on a motion vector difference (MVD) in units of integer samples; deriving a first motion vector predictor (MVP) in units of fraction samples on the basis of neighboring blocks of a current block; deriving a second MVP in units of integer samples on the basis of the first MVP; determining a first motion vector (MV) in units of integer samples for the current block on the basis of the second MVP and the MVD; determining a second MV in units of fraction samples for the current block on the basis of the first MVP and the first MV; and generating a prediction sample for the current block on the basis of the second MV. The present invention is capable of increasing the accuracy of the MV while reducing the amount of data allocated to the MVD.