Non-Separable Primary Transforms for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards, such as those prior to ITU-T H.265/HEVC, rely on fixed separable transforms which limit coding efficiency, and while HEVC introduced additional transforms like DST-7, they still do not fully exploit the potential of non-separable transforms for improved compression.
Innovation Solution
The implementation of non-separable primary transforms (NSPTs) directly on residual data, without the need for separable transforms, to enhance coding efficiency by applying dedicated kernels for various block shapes, thereby improving compression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fixed separable transforms are used, then device complexity is reduced, but coding efficiency deteriorates
Solution Approach 1:
The transform process is segmented into separable horizontal and vertical transforms, which simplifies the computational structure and reduces device complexity while maintaining coding efficiency through the use of dedicated kernels for different block shapes and sizes
Solution Approach 2:
The patent introduces adaptive transform selection where the transform type is dynamically chosen based on block shape, size, and prediction mode, allowing the system to optimize coding efficiency for different content characteristics while maintaining a manageable device complexity through a finite set of transform kernels
2Productivity
If non-separable primary transforms are used, then coding efficiency is improved, but device complexity increases
Solution Approach 1:
Different transform kernels are applied to different regions (blocks) based on their local characteristics such as shape, size, and prediction mode, allowing optimal transform selection for each block while maintaining overall system manageability through a structured set of kernels
Solution Approach 2:
The patent changes transform parameters (kernel selection) based on block characteristics and prediction modes, enabling adaptive optimization of coding efficiency without requiring a complete redesign of the transform architecture, thus balancing complexity and performance
3Productivity
If larger transform kernels are used, then compression performance is improved, but computational complexity increases
Solution Approach 1:
The patent applies larger transform kernels selectively only when beneficial for compression performance, rather than universally, thus achieving improved compression where needed while avoiding unnecessary computational complexity in cases where smaller kernels suffice
Data Source
AI summary
An example device for decoding video data includes a memory configured to store video data; and a processing system comprising one or more processors implemented in circuitry, the processing system being configured to: inverse transform a block of transform coefficients of a block of the video data using an inverse non-separable primary transform (NSPT), without using an inverse separable transform, to reconstruct a residual block of the block of the video data; and decode the block using the residual block.