LFNST Index Parsing for Transform Coefficient Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, such as 4K and 8K ultra high definition, leads to higher transmission and storage costs due to increased bit amounts, and existing compression techniques are inefficient for immersive media like virtual reality and hologram content.
Innovation Solution
The proposed method involves applying the Low-Frequency Non-Separable Transform (LFNST) to transform coefficients, which includes deriving modified transform coefficients by parsing an LFNST index based on a variable indicating the presence of significant coefficients outside the DC component, and using a transform skip flag to optimize the coding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image compression techniques are used for high-resolution images, then transmission and storage costs increase due to increased bit amounts, but image quality and resolution requirements cannot be met
Solution Approach 1:
The patent applies Low-Frequency Non-Separable Transform (LFNST) to transform coefficients, changing the transformation parameters and methods to achieve better compression efficiency. By modifying the transform domain and applying LFNST specifically to low-frequency components, the patent optimizes the distribution of transform coefficients, enabling more efficient quantization and coding while maintaining image quality.
2Productivity
If existing compression techniques are applied to immersive media, then coding efficiency is insufficient, but the complexity of handling various image features increases
Solution Approach 1:
The patent segments the transform coefficient processing into different frequency bands, applying LFNST specifically to low-frequency components while handling high-frequency components differently. This segmentation allows the patent to optimize compression for the most important visual information (low-frequency) without unnecessarily complicating the processing of less important high-frequency components.
Solution Approach 2:
The patent applies different processing strategies to different frequency regions based on their importance. Low-frequency transform coefficients, which contain the most significant visual information, receive special treatment through LFNST application, while high-frequency coefficients are processed using conventional methods. This local quality approach optimizes overall coding efficiency without uniformly increasing complexity across all frequency bands.
3Productivity
If transform coefficients are processed without optimization, then LFNST index coding efficiency is low, but additional processing steps increase computational complexity
Solution Approach 1:
The patent performs preliminary organization of transform coefficients by frequency bands before LFNST index coding. By pre-grouping and identifying significant low-frequency coefficients, the patent prepares the data structure in advance to enable more efficient index coding. This preliminary action reduces the computational burden during the actual LFNST application and indexing stages.
Data Source
AI summary
An image decoding method according to the present document comprises the step of deriving modified transform coefficients by applying LFNST to transform coefficients, wherein the step of deriving the modified transform coefficients comprises the steps of: deriving a variable indicating whether an effective coefficient is present in a DC component of a current block; and parsing an LFNST index on the basis of whether the variable indicates that the effective coefficient is present in a position other than the DC component, wherein the variable may be derived on the basis of an individual transform skip flag value for a color component of the current block.