Image Coding Using LFNST and MTS Indices
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 bit rates, resulting in increased transmission and storage costs. Additionally, the need for efficient compression techniques is exacerbated by the growing interest in immersive media like virtual and augmented reality.
Innovation Solution
The proposed solution involves an image coding method and apparatus that utilize Low-Frequency Non-Separable Transform (LFNST) and Multiple Transform Selection (MTS) to enhance coding efficiency. This method includes residual coding, deriving residual samples using LFNST or MTS, and generating a reconstructed picture based on these samples, with the LFNST and MTS indices signaled at the coding unit level.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images/videos (4K, 8K) are transmitted or stored using conventional methods, then image quality is improved, but transmission cost and storage cost increase due to higher bit rates
Solution Approach 1:
The patent applies Low-Frequency Non-Separable Transform (LFNST) and Multiple Transform Selection (MTS) to change the transform parameters and types used in video coding. By selecting different transform kernels (e.g., DCT, DST, DRT) based on prediction mode and block characteristics, the patent achieves better energy compaction and coding efficiency, thereby reducing bit rate while maintaining high image quality
Solution Approach 2:
The patent divides the transform process into multiple stages: primary transform followed by secondary transform (LFNST). This segmentation allows different transform types to be applied to different frequency components and block regions, optimizing compression for each segment while preserving overall image quality
2Productivity
If conventional transform methods are used for video coding, then device complexity is kept simple, but coding efficiency is insufficient for high-resolution videos
Solution Approach 1:
The patent introduces dynamic transform selection where the transform type is adaptively chosen based on prediction mode, block size, and other coding conditions. The MTS index and LFNST index are dynamically determined during encoding, allowing the system to optimize coding efficiency for each specific situation while managing complexity through conditional logic
Solution Approach 2:
The patent performs preliminary classification of transform types based on prediction mode and block characteristics before actual transform processing. By determining the appropriate transform kernel in advance (through index selection), the patent reduces computational complexity during the actual transform operation while maintaining high coding efficiency
Data Source
AI summary
An image decoding method according to the present document comprises: a residual coding step for parsing residual information received in a residual coding level and arranging transform coefficients for a current block according to a predetermined scanning order; a step for deriving residual samples by applying at least one of LFNST or MTS to the transform coefficients; and a step for generating a reconstructed picture on the basis of the residual samples, wherein the LFNST is performed on the basis of an LFNST index indicating an LFNST kernel, the MTS is performed on the basis of an MTS index indicating an MTS kernel, the LFNST index and the MTS index are signaled in a coding unit level, and the MTS index is signaled immediately after the signaling of the LFNST index.