Selective Block Motion Compensation for Lower Decoding Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image decoding technologies using overlapped block motion compensation (OBMC) face increased memory bandwidth and calculation requirements due to its application without considering block sizes and prediction directions, leading to reduced encoding performance.
Innovation Solution
An image decoding device that applies OBMC based on specific conditions such as block sizes, prediction directions, and types of adjacent blocks, including determining whether to apply OBMC, changing prediction methods, and adjusting weight coefficients to optimize memory usage and calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OBMC is applied regardless of block size or prediction direction, then prediction accuracy is improved, but memory bandwidth and calculation requirements increase
Solution Approach 1:
The patent changes the application parameters of OBMC by introducing determination conditions based on block size and prediction direction type. The system dynamically adjusts whether to apply OBMC or not, and when to apply it, based on these parameters. This resolves the contradiction by avoiding unnecessary OBMC applications that increase memory bandwidth while preserving accuracy where beneficial.
Solution Approach 2:
The patent makes the OBMC application dynamic rather than static. Instead of applying OBMC uniformly to all blocks, the system dynamically determines applicability based on real-time analysis of block size and prediction direction. This dynamic approach allows the system to adapt to different coding scenarios and optimize the balance between prediction accuracy and memory bandwidth usage.
2Measurement precision
If OBMC is applied regardless of block size or prediction direction, then prediction accuracy is improved, but calculation requirements increase
Solution Approach 1:
The patent changes the application parameters of OBMC by introducing determination conditions based on block size and prediction direction type. The system dynamically adjusts whether to apply OBMC or not, and when to apply it, based on these parameters. This resolves the contradiction by avoiding unnecessary OBMC applications that increase calculation requirements while preserving accuracy where beneficial.
Solution Approach 2:
The patent applies OBMC partially rather than excessively. Instead of uniformly applying OBMC to all blocks regardless of suitability, the system selectively applies it only to blocks that meet the determination conditions (specific block sizes and prediction direction combinations). This partial application reduces unnecessary calculations while maintaining accuracy improvements where OBMC is truly beneficial.
3Loss of energy
If OBMC is applied to reduce memory bandwidth, then encoding performance improves, but decoding complexity increases
Solution Approach 1:
The patent changes the application parameters of OBMC by introducing determination conditions based on block size and prediction direction type. The system dynamically adjusts whether to apply OBMC or not, and when to apply it, based on these parameters. This resolves the contradiction by avoiding unnecessary OBMC applications that increase memory bandwidth while preserving accuracy where beneficial.
4Use of energy by moving object
If OBMC is selectively applied based on determination conditions, then memory bandwidth and calculation requirements are reduced, but prediction accuracy may deteriorate
Solution Approach 1:
The patent applies local quality by treating different blocks differently based on their specific characteristics (size and prediction direction). Instead of a uniform approach, the system identifies local conditions where OBMC is beneficial and applies it only there. This resolves the contradiction by concentrating OBMC applications in regions where they provide maximum benefit while avoiding them where they would be detrimental, thus optimizing both memory bandwidth and prediction accuracy.
Data Source
AI summary
An image decoding device includes: a motion compensation predictor configured to generate a prediction image signal of a prediction target block based on information about a motion vector in a unit of the block and a reference frame; an overlapped block motion compensator carrying out an overlapped block motion compensation process of correcting the prediction image signal of the prediction target block by subjecting the prediction image signal of the prediction target block and a prediction image signal generated based on information about a motion vector and a reference frame of an adjacent block of the prediction target block to weighted averaging; and a determinator determining whether the overlapped block motion compensation process is applied to the prediction target block or not. The determinator determines application of the overlapped block motion compensation process to the prediction target block, when unidirectional prediction is applied to the adjacent block.


