Merge-Based Motion Vector Offsets for Video Compression Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-definition video services has led to significant challenges in data volume, with existing video compression standards like HEVC showing limitations in efficiency.
Innovation Solution
A method of refining motion vectors using offset vectors in video encoding and decoding, with adaptive determination of magnitude and direction based on index information, and signaling these vectors through picture parameter sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If HEVC compression standard is used to reduce data volume, then compression performance is improved, but compression efficiency is insufficient for high-definition video services
Solution Approach 1:
The patent refines motion vectors by introducing offset vectors that adjust the magnitude and direction parameters of motion information. This parameter refinement allows for more precise motion modeling, thereby improving compression efficiency while maintaining reduced data volume requirements for high-definition video services
2Productivity
If motion vectors are refined using offset vectors, then inter prediction efficiency is improved, but device complexity increases
Solution Approach 1:
The patent segments the motion vector refinement process into distinct components: obtaining a motion vector from merge candidates and adding an offset vector to refine it. This segmentation allows the complex refinement operation to be broken down into manageable steps that can be efficiently implemented in encoding and decoding devices
Solution Approach 2:
The offset vector acts as an intermediary element that bridges the motion vector from merge candidates and the final refined motion vector. By introducing this intermediate representation, the patent simplifies the overall complexity of motion vector refinement while maintaining improved inter prediction efficiency
Data Source
AI summary
A video decoding method according to the present disclosure includes the steps that: a merge candidate list for a current block is generated; a merge candidate for the current block is determined among merge candidates included in the merge candidate list; an offset vector for the current block is derived; and a motion vector for the current block is derived by adding the offset vector to a motion vector of the merge candidate.