LFNST Index Parsing for Image Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, such as 4K and 8K UHD, leads to higher transmission and storage costs due to increased data amounts, and there is a need for efficient compression techniques to handle immersive media formats like VR and AR, which require advanced image/video compression methods.
Innovation Solution
The implementation of an image coding method using Low-Frequency Non-Separable Transform (LFNST) to derive modified transform coefficients and encode residual information, where the LFNST index is parsed based on block dimensions, tree-type, color format, and the application of an Image Signal Processor (ISP), optimizing transform index coding and compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution and high-quality images/videos (4K, 8K UHD) are transmitted or stored, then image quality and resolution are improved, but transmission cost and storage cost increase due to increased data amount
Solution Approach 1:
The image block is divided into multiple sub-partition transform blocks, and LFNST is selectively applied to specific sub-partitions based on their characteristics. This segmentation allows different compression strategies to be applied to different regions, achieving better overall compression efficiency while maintaining image quality.
Solution Approach 2:
LFNST is applied selectively to specific sub-partition transform blocks based on local characteristics such as block size, transform type, and content features. This local quality approach ensures that compression is optimized for each region without compromising overall image quality.
2Productivity
If conventional compression techniques are used for immersive media (VR, AR), then basic compression is achieved, but coding efficiency is insufficient for high-resolution and various features
Solution Approach 1:
The LFNST application is dynamically determined based on multiple factors including block size, transform type (DST VII, DCT II), color format (4:2:0, 4:4:4), and ISP application status. This dynamic adaptation allows the coding system to optimize for each specific scenario, achieving high compression efficiency without excessive complexity.
Solution Approach 2:
The patent changes multiple parameters simultaneously to optimize compression: block dimension thresholds (4x4, 8x8, 16x16), transform type selection, color format handling, and ISP application status. These parameter changes enable adaptive compression that achieves high productivity for immersive media.
3Productivity
If transform index coding is applied to all blocks, then compression is achieved, but coding efficiency is reduced due to unnecessary indexing in blocks where LFNST should not be applied
Solution Approach 1:
The LFNST index parsing is extracted and separated from universal application. Instead of applying LFNST to all blocks, the patent extracts the decision logic to selectively parse LFNST index only for blocks that meet specific criteria (size, transform type, color format, ISP status), eliminating unnecessary indexing and improving coding efficiency.
Solution Approach 2:
Rather than applying LFNST universally (excessive action), the patent applies it partially only to sub-partition transform blocks that meet specific conditions. This partial application avoids unnecessary coding overhead while maintaining compression efficiency where it is most beneficial.
Data Source
AI summary
An image decoding method according to the present document comprises a step for deriving a modified transform coefficient, wherein the step for deriving the modified transform coefficient comprises a step for determining whether or not to parse an LFNST index on the basis of whether or not the width and height of a current block satisfy a condition about whether the LFNST can be applied, and whether or not the condition about whether the LFNST can be applied is satisfied is determined on the basis of a tree type and a color format of the current block and whether or not an ISP is applied to the current block.