Weighted Prediction Tables for Lower-Overhead Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution, high-quality images and videos, particularly in virtual reality and augmented reality, necessitates a more efficient compression technique to reduce transmission and storage costs while effectively handling diverse image/video characteristics.
Innovation Solution
A method and apparatus for enhancing video/image coding efficiency by efficiently signaling a prediction weighted syntax table, reducing signaling overhead, and improving compression through techniques like intra prediction, inter prediction, and residual processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high resolution and high quality image/video are transmitted, then image quality is improved, but transmission cost and storage cost increase
Solution Approach 1:
The patent extracts only the essential visual information by performing transform coding on image blocks, converting spatial domain data to frequency domain coefficients. Only significant coefficients are transmitted and stored, separating essential visual information from redundant data, thereby reducing transmission and storage costs while maintaining image quality.
Solution Approach 2:
The patent applies quantization to transform coefficients, changing their precision parameters based on importance. Less important coefficients are quantized with coarser precision while critical coefficients retain finer precision, enabling quality preservation with reduced data volume for cost-effective transmission and storage.
2Measurement precision
If prediction weighted table syntax is signaled explicitly, then prediction accuracy is improved, but signaling overhead increases
Solution Approach 1:
The patent performs preliminary action by deriving prediction weighted tables from previously signaled reference picture lists and motion information. Instead of signaling complete weighted tables, the system prepares and stores prediction weights in advance based on available reference data, reducing signaling overhead while maintaining prediction accuracy through pre-computed weight values.
Solution Approach 2:
The patent introduces an intermediary mechanism where prediction weights are derived as intermediate values from reference picture lists and motion compensation parameters. Rather than directly signaling final weighted prediction values, the system uses intermediate derivation steps that leverage existing decoded information, reducing the amount of new information that must be transmitted.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A video decoding method executed by a video decoding apparatus according to the present document comprises the steps of: parsing a flag associated with weighted prediction from a bitstream; on the basis of the flag, parsing a prediction weighted table syntax from the bitstream; and on the basis of the prediction weighted table syntax, executing a weighted prediction on a current block in a current picture to reconstruct the current picture, wherein on the basis of the value of the flag being 1, the prediction weighted table syntax may be parsed from a picture header of the bitstream, and on the basis of the value of the flag being 0, the prediction weighted table syntax may be parsed from a slice header of the bitstream.