Bi-Directional Optical Flow Prediction with One Motion Vector
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image encoding/decoding methods using bi-directional optical flow (BIO) are limited to cases where two pieces of motion information are available, and there is a need for techniques to apply BIO when only one piece of motion information is provided, along with methods to reduce complexity and variability in calculating BIO offsets.
Innovation Solution
The method involves determining whether to derive second motion information based on first motion information, calculating BIO offsets in a sub-group unit, and performing BIO prediction based on both motion information, with the option to apply BIO in a block unit, allowing for variable sub-group sizes and additional processing steps like deblocking filtering and inverse-transform in the sub-group unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If BIO is applied only when two pieces of motion information are provided, then the decoding complexity is reduced, but the coding efficiency is limited
Solution Approach 1:
The patent changes the parameter of motion information availability by deriving a second motion vector from the first motion vector when only one piece of motion information is provided. This allows BIO to be applied in cases where traditionally only one motion vector was available, thereby improving coding efficiency without requiring two separate motion information sources
Solution Approach 2:
The patent introduces an intermediary derivation process that generates a second motion vector from the first motion vector. This intermediary step enables the BIO algorithm to function with a single provided motion vector by creating a synthetic second vector, thus resolving the contradiction between limited motion information availability and BIO application requirements
2Productivity
If BIO is applied in block unit, then the coding efficiency is improved, but the computational complexity increases
Solution Approach 1:
The patent segments the block into sub-groups for calculating BIO offsets. By dividing the block into smaller sub-groups, the computational complexity is managed more effectively while still maintaining the ability to apply BIO at the block level for improved coding efficiency
Solution Approach 2:
The patent applies BIO selectively by determining whether to derive second motion information based on the first motion vector. This partial application approach allows BIO to be used where it provides the most benefit while avoiding unnecessary computational complexity in cases where it would provide minimal improvement
3Ease of manufacture
If fixed sub-group size is used for calculating BIO offset, then the processing is simplified, but the adaptability to different block sizes is reduced
Solution Approach 1:
The patent introduces dynamic sub-group sizing where the sub-group size for calculating BIO offsets can vary based on the block characteristics. This dynamic approach allows the system to adapt to different block sizes and configurations while maintaining reasonable processing complexity through structured size variations
Data Source
AI summary
An image encoding/decoding method is disclosed. An image decoding method of the present invention may comprises determining step for determining whether or not to derive second motion information when a current block includes first motion information only and for which bi-directional prediction is available, a deriving step for deriving the second motion information based on the first motion information, and a predicting step for generating a prediction block of the current block by performing BIO prediction based on the first motion information and the second motion information.