Image Coding With Adaptive Merge Candidates for Motion Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies struggle to efficiently encode and decode high-resolution and stereoscopic image content, particularly in handling motion vectors and intra/inter predictions.
Innovation Solution
The method and device adaptively partition pictures, utilize a merge candidate list for motion information, and apply a motion vector difference value (MVD) to enhance encoding/decoding efficiency, incorporating spatial and temporal merge candidates with adaptive candidate selection and scaling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video compression technology is used for high-resolution and stereoscopic image content, then encoding complexity increases and compression efficiency decreases
Solution Approach 1:
The current block is divided into multiple sub-blocks for independent motion compensation processing. This segmentation allows the encoder to handle complex high-resolution content by breaking it into smaller, more manageable units, improving compression efficiency without overwhelming the encoding system with monolithic complexity.
Solution Approach 2:
Different motion compensation modes (merge mode and AMVP mode) are selectively applied to different sub-blocks based on their local characteristics. This local quality approach optimizes compression efficiency by using the most appropriate mode for each region while managing overall encoding complexity through systematic mode selection.
2Measurement precision
If motion vector precision is increased for accurate motion compensation, then encoding bitrate increases
Solution Approach 1:
The patent employs two different motion vector precision parameters: 1/4-pel precision for merge mode and 1-pel precision for AMVP mode. This parameter change strategy allows the system to achieve accurate motion compensation where needed while controlling bitrate by using coarser precision where appropriate, thus resolving the contradiction between accuracy and bitrate.
Solution Approach 2:
The motion vector precision is dynamically adjusted based on the prediction mode and block characteristics. The system transitions between different precision levels (1/4-pel and 1-pel) depending on the specific encoding context, allowing optimal balance between motion vector accuracy and encoding bitrate for each situation.
3Measurement precision
If multiple merge candidates are considered for motion prediction, then prediction accuracy improves but processing time increases
Solution Approach 1:
The patent limits the merge candidate list to a specific number of candidates (typically 4 or 5) rather than considering all possible candidates. This partial action approach provides sufficient prediction accuracy for most cases while significantly reducing processing time compared to exhaustive candidate evaluation, effectively resolving the time-accuracy trade-off.
Solution Approach 2:
Motion information from previously decoded blocks is stored and made available as merge candidates before the current block is processed. This preliminary action of pre-computing and storing motion information from neighboring blocks reduces the processing time for the current block while maintaining prediction accuracy through informed candidate selection.
Data Source
AI summary
The image encoding/decoding method and device according to the present invention can construct a merge candidate list of a current block, derive motion information of the current block on the basis of the merge candidate list and a merge index of the current block, derive a motion vector difference of the current block, correct a motion vector of the current block by means of the motion vector difference, and carry out motion compensation on the current block by means of the corrected motion vector.