Image Coding Merge Mode Using Reference Picture Motion Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image coding methods in the merge mode are limited to adjacent blocks within the current picture, leading to decreased coding efficiency when the adjacent block lacks a motion vector, such as in intra-coded blocks.
Innovation Solution
The proposed method extends the merge process to include coding result information from reference pictures different from the current picture by calculating and utilizing alternative reference indices and motion vectors, and adding a flag to the bitstream to indicate the use of these values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the merge mode is limited to adjacent blocks within the current picture, then the device complexity is reduced, but the coding efficiency deteriorates when adjacent blocks are intra-coded
Solution Approach 1:
The patent extends the merge candidate search from the spatial dimension (adjacent blocks in current picture) to the temporal dimension (corresponding blocks in reference pictures). This allows the system to access motion vectors from previously decoded pictures, thereby maintaining coding efficiency even when adjacent blocks are intra-coded without compromising device complexity
2Productivity
If coding result information from reference pictures is incorporated into the merge process, then the coding efficiency is improved, but the device complexity increases
Solution Approach 1:
The patent performs preliminary decoding and storage of reference picture information during the reference picture decoding phase. By pre-processing and storing motion vector data from reference pictures, the system avoids complex real-time processing during the current picture decoding, thus improving coding efficiency while controlling device complexity
Solution Approach 2:
The patent copies motion vector information from corresponding blocks in reference pictures to the current block's merge candidates. This copying mechanism allows efficient reuse of previously decoded motion data without requiring complex recalculation, thereby improving coding efficiency while maintaining manageable device complexity
Data Source
AI summary
An image coding method of coding a current block to be coded using a first reference index indicating a first reference picture and a first motion vector, includes: calculating, using a second reference index and a second motion vector used to code a corresponding block, a third reference index and a third motion vector as respective candidates for the first reference index and the first motion vector; determining a value of a flag that indicates whether or not to code the current block using the third reference index and the third motion vector as the first reference index and the first motion vector; and coding the current block using the first reference index and the first motion vector according to the value of the flag, and adding the value of the flag to a bitstream generated by the coding.