Image Transform Coding With Block-Limited Non-Separable Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images/videos, including immersive media, necessitates a highly efficient image/video compression technique to reduce transmission and storage costs while effectively handling diverse image features.
Innovation Solution
An image coding method and apparatus that employs a multiple transform technique, including non-separable transforms, to optimize block sizes and enhance transform index coding efficiency, with specific conditions for applying inverse non-separable transforms based on block sizes and color formats.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If non-separable transforms are applied to large blocks, then transform efficiency is improved, but computational complexity increases excessively
Solution Approach 1:
The patent segments the transform processing by applying non-separable transforms only to specific sub-blocks (e.g., 4×4 or 8×8 regions) within a larger transform unit, rather than applying the computationally intensive non-separable transform to the entire large block. This segmentation maintains transform efficiency for important regions while reducing overall computational complexity.
Solution Approach 2:
The patent applies different transform types to different regions based on local characteristics. Non-separable transforms are applied selectively to regions where they provide the most benefit (e.g., regions with specific directional characteristics), while other regions use standard separable transforms. This local quality approach optimizes transform efficiency where needed while controlling overall computational complexity.
2Adaptability or versatility
If multiple transform types are supported, then coding flexibility is improved, but device complexity increases
Solution Approach 1:
The patent implements dynamic transform selection where the transform type is chosen based on block characteristics, prediction mode, and other contextual factors. The system can adaptively switch between separable and non-separable transforms, and between different non-separable transform kernels, providing coding flexibility while managing device complexity through conditional logic rather than always supporting all transform types simultaneously.
Solution Approach 2:
The patent supports multiple transform types but applies them partially - only to specific blocks or regions where they are most beneficial. Rather than implementing all transform types for all blocks (excessive action), the system selectively applies non-separable transforms to a subset of blocks, providing the flexibility benefits where needed while reducing the overall device complexity burden.
3Measurement precision
If transform index coding is enhanced, then coding precision is improved, but bit rate increases
Solution Approach 1:
The patent enhances transform index coding precision but only for the subset of blocks where non-separable transforms are applied. Rather than using detailed transform indexing for all blocks (which would increase bit rate excessively), the system applies enhanced indexing selectively to blocks benefiting from non-separable transforms, maintaining coding precision where needed while controlling overall bit rate.
Solution Approach 2:
The patent segments the transform indexing strategy by using different indexing approaches for different block types. For blocks with non-separable transforms, more precise transform indexing is used to accurately specify the transform kernel. For other blocks, standard or simplified indexing is used. This segmentation maintains coding precision for important cases while reducing overall bit rate.
Data Source
AI summary
including: a step for receiving a bitstream including residual information; a step for deriving transform coefficients for a target block on the basis of the residual information; a step for deriving corrected transform coefficients on the basis of an inverse non-separable transform of the transform coefficients; and a step for deriving residual samples for the target block on the basis of an inverse primary transform of the corrected transform coefficients, wherein the inverse non-separable transform is performed when the size of the target block is equal to or smaller than the size of a prescribed maximum transform application block.