Transform-Based Image Coding with Conditional LFNST Signaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, particularly in immersive media formats, leads to higher transmission and storage costs due to increased bit amounts, necessitating a more efficient image/video compression technique.
Innovation Solution
An image coding method and apparatus utilizing LFNST (Low Frequency Non-Separable Transform) to enhance compression efficiency by zeroing out specific regions and signaling an LFNST index when significant coefficients exist outside the DC region, along with efficient transform index coding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images/videos are transmitted or stored, then image quality is improved, but transmission cost and storage cost increase due to increased bit amount
Solution Approach 1:
The patent segments the transform coefficient block into multiple regions (first region at top-left, second region, third region) and applies different coding strategies to each region. The first region is coded using LFNST transform index, while other regions use conventional coding, allowing optimized compression for different frequency components without sacrificing overall image quality
Solution Approach 2:
The patent applies local quality by treating different regions of the transform coefficient block differently based on their frequency characteristics. The first region (top-left) containing low-frequency coefficients receives specialized LFNST coding attention, while higher frequency regions are handled with standard methods, optimizing the balance between compression efficiency and quality preservation
2Adaptability or versatility
If LFNST transform index is parsed for all blocks, then transform coding flexibility is improved, but coding complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing LFNST transform matrices for different block sizes (4x4, 8x8, 16x16) and frequency bands before the actual coding process. This allows the decoder to quickly retrieve and apply the appropriate transform without complex real-time calculations, reducing decoding complexity while maintaining flexibility
Solution Approach 2:
The patent implements dynamics by making the LFNST transform index parsing conditional rather than mandatory. The index is only parsed and applied when certain conditions are met (e.g., when the block size and frequency band require it), allowing the system to adaptively choose between LFNST coding and conventional coding based on the specific block characteristics, thus balancing flexibility with complexity
Data Source
AI summary
An image decoding method according to the present document may comprise the steps of: deriving a first variable indicating whether there is a valid coefficient in a region excluding a DC region from a current block; deriving a second variable indicating whether there is a valid coefficient in a second region excluding a first region formed at the upper left end of the current block; when the first variable indicates that the valid coefficient exists in the region excluding the DC region, and the second variable indicates that the valid coefficient does not exist in the second region, parsing an LFNST index from the bitstream; and applying an LFNST matrix derived on the basis of the LFNST index to transform coefficients in the first region, to derive the modified transform coefficients.