Video Residual Decoding With Non-Separable Primary Transforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques face limitations in achieving optimal compression efficiency due to the reliance on separable transforms, which can be improved by incorporating non-separable primary transforms (NSPTs) that directly process residual data without requiring separable transforms beforehand.
Innovation Solution
Implementing non-separable primary transforms (NSPTs) during video encoding and decoding processes to directly process residual blocks, bypassing the need for separable transforms, thereby enhancing coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If separable transforms are used for video coding, then the implementation complexity is reduced and processing is simplified, but the compression efficiency is limited and cannot achieve optimal performance
Solution Approach 1:
The patent segments the transform process into two distinct stages: a separable primary transform (SPT) that provides computational efficiency, and a non-separable secondary transform (NSST) that enhances compression performance. This segmentation allows each transform type to operate optimally in its designated stage, resolving the contradiction between implementation simplicity and compression efficiency.
Solution Approach 2:
The separable primary transform is applied first to the residual block before the non-separable secondary transform. This preliminary action reduces the data complexity and energy distribution, creating favorable conditions for the subsequent NSST to achieve better compression without excessive computational burden.
2Productivity
If non-separable primary transforms are applied directly to residual data, then compression efficiency is improved, but the computational complexity increases compared to separable transforms
Solution Approach 1:
The transform process is divided into a first transform stage (separable) and a second transform stage (non-separable). This segmentation enables the system to leverage the computational advantages of separable transforms while still incorporating the compression benefits of non-separable transforms, thus balancing complexity and efficiency.
Solution Approach 2:
The non-separable transform is applied partially, specifically only to the transform coefficients produced by the separable primary transform, rather than directly to all residual data. This partial application reduces the overall computational burden while still achieving improved compression efficiency in critical regions.
3Productivity
If larger transform kernels are used in NSPT, then compression performance is enhanced, but the processing load and computational resources required increase
Solution Approach 1:
The use of larger transform kernels is segmented into the second transform stage only, after the initial separable transform has reduced the data dimensionality and concentrated energy. This allows larger kernels to be used effectively without proportionally increasing the overall computational load across the entire processing pipeline.
Solution Approach 2:
The separable primary transform with smaller kernels is applied first to prepare the data, concentrating energy into fewer coefficients. This preliminary action reduces the effective processing load for subsequent larger transform kernels, as they operate on already-compressed, lower-dimensional data.
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.