Quantization Matrix Selection Using a Single Block Identifier
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image encoding/decoding methods face limitations in improving both objective and subjective image quality due to the lack of consideration of factors such as prediction mode, color component, size, form, one-dimensional transform type, and two-dimensional transform combination.
Innovation Solution
The method involves decoding and encoding information on a quantization matrix based on a current block's size, prediction mode, or color component, adjusting coefficient values, and encoding/decoding this information using a single identifier, which includes determining and using a quantization matrix considering these factors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional quantization matrix encoding/decoding is used considering only prediction mode, color component, or size, then the encoding process is simple, but the objective image quality and subjective image quality are limited
Solution Approach 1:
The patent applies parameter changes by considering multiple factors (prediction mode, color component, size, form, transform type, and transform combination) to determine the quantization matrix, rather than relying on a single parameter. This multi-parameter approach enables more precise control over the quantization process, improving both objective and subjective image quality while managing encoding complexity through systematic parameter integration.
2Manufacturing precision
If multiple factors are considered for quantization matrix selection, then image quality improves, but the encoding complexity increases
Solution Approach 1:
The patent segments the quantization matrix selection process into distinct components: prediction mode identification, color component determination, size classification, form identification, transform type selection, and transform combination determination. By dividing the complex selection process into manageable segments, the system can evaluate multiple factors systematically without overwhelming encoding complexity, allowing high-quality image reconstruction through structured multi-factor analysis.
3Productivity
If a single identifier is used to encode multiple factors, then the bitstream efficiency improves, but the information representation becomes more complex
Solution Approach 1:
The patent merges multiple factors (prediction mode, color component, size, form, transform type, and transform combination) into a single identifier that is transmitted in the bitstream. This consolidation reduces the overall bitstream size and improves encoding efficiency by eliminating the need to separately encode each factor. The single identifier integrates all necessary information, allowing the decoder to reconstruct the quantization matrix by interpreting the combined parameters without requiring separate signaling for each factor.
Data Source
AI summary
Disclosed herein is a method of encoding/decoding an image. The image decoding method includes decoding information on a quantization matrix from a bitstream, acquiring a quantization matrix of a current block based on the information on the quantization matrix, and dequantizing the current block using the quantization matrix of the current block. The acquiring of the quantization matrix of the current block includes deriving a single identifier using at least one of a size of the current block, a prediction mode or a color component, and acquiring the quantization matrix of the current block based on the single identifier.