Weighted Prediction and Bi-Directional Optical Flow for Image Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need for high-efficient image compression technology to effectively transmit, store, and reproduce high-resolution and high-quality images, as the increased amount of transmitted information leads to higher transmission and storage costs.
Innovation Solution
An image encoding/decoding method and apparatus that performs weighted prediction, including bi-prediction with CU-level weight (BCW), in consideration of prediction refinement with optical flow (PROF), to improve encoding/decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images are transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent applies parameter changes by implementing weighted prediction and bi-directional optical flow techniques that modify how pixel values are predicted and encoded. By changing the prediction parameters and using weighted combinations of reference blocks, the system achieves better compression efficiency while maintaining high image quality, thus reducing the quantity of transmitted information without sacrificing manufacturing precision (image quality).
2Quantity of substance
If the amount of transmitted information is reduced, then transmission cost is reduced, but image quality deteriorates
Solution Approach 1:
The patent employs feedback mechanisms through bi-directional optical flow that uses both forward and backward reference pictures to predict current block motion. This feedback approach allows the system to maintain accurate motion compensation with reduced data transmission, thereby preserving image quality while reducing the amount of transmitted information through more efficient prediction.
3Productivity
If weighted prediction with CU-level weight is performed, then encoding efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the current block into multiple coding units (CUs) and applying different weight indices to each CU level. This segmentation approach allows weighted prediction to be performed efficiently at appropriate granularities, improving encoding efficiency while managing device complexity through hierarchical processing rather than uniform complex operations across the entire block.
Data Source
AI summary
An image encoding/decoding method and apparatus are provided. An image decoding method according to the present disclosure is performed by an image decoding apparatus. The image decoding method comprises deriving a first flag specifying whether to perform weighted prediction on a current block and a weight index BcwIdx of bi-prediction with CU-level weight (BCW) for the current block, determining whether to perform default weighted prediction or explicit weighted prediction on the current block based on the first flag and BcwIdx, and generating a prediction block for the current block by performing the determined method.