Selective Secondary Transform for Video Coding Complexity Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video coding schemes face increased processing complexity due to secondary transforms, which concentrate energy in low-frequency components but require significant computational resources, especially for long transform components.
Innovation Solution
An image encoding and decoding device that performs transforms on coding units with a first transformer and a second transformer, which applies transform on specific regions of varying sizes or shapes, reducing processing complexity while maintaining coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If secondary transform is performed on all transform coefficients to concentrate energy in low-frequency components, then coding efficiency is improved, but processing complexity and computational resources increase significantly
Solution Approach 1:
The patent applies secondary transform selectively to specific regions (e.g., blocks with certain characteristics like presence of DC coefficient or specific coefficient patterns) rather than uniformly to all transform coefficients. This localized application maintains coding efficiency for regions that benefit from energy concentration while avoiding unnecessary processing in regions where secondary transform provides minimal gain, thus resolving the contradiction between coding efficiency and processing complexity.
Solution Approach 2:
The patent performs secondary transform on only a portion of the transform coefficients based on predetermined conditions (such as block type, coefficient distribution, or region of interest), rather than applying it excessively to all coefficients. This partial application achieves sufficient energy concentration for effective coding while significantly reducing the computational burden compared to universal application.
2Manufacturing precision
If non-separable transform is used for secondary transform to increase energy concentration of diagonal direction components, then line quality in diagonal direction is improved, but transform complexity increases due to O(N^2) or O(NlogN) operation amount
Solution Approach 1:
The patent applies non-separable secondary transform selectively to specific regions where diagonal line structures are present or expected, rather than applying it uniformly across the entire block. This targeted application improves line quality in diagonal directions where needed while avoiding the high computational cost of non-separable transform in regions where it provides no benefit.
Solution Approach 2:
The patent performs non-separable secondary transform on only a subset of transform coefficients (e.g., low-frequency components or specific frequency bands) rather than applying it to all coefficients. This partial application maintains the ability to capture diagonal line energy concentration while reducing the overall transform complexity from O(N^2) to a lower computational burden.
Data Source
AI summary
A device is provided with: a first transformer which transforms an coding unit (CU); and a second transformer which transforms a part of first transform coefficients output from the first transformer, wherein the second transformer transforms at least any of the first transform coefficients for a region (first region) having different sizes in a horizontal direction and a vertical direction or the first transform coefficients for a non-rectangular region (second region).