Multiple Transform Selection for Video Coding Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face limitations in coding efficiency due to the fixed use of separable transforms, which do not adequately accommodate the varying statistics of residual samples, leading to suboptimal compression performance.
Innovation Solution
The implementation of multiple transform selection (MTS) techniques, including the use of secondary transforms in addition to primary transforms, and simplified Low-Frequency Non-separable Transformations (LFNST), allows for adaptive transform selection to improve coding efficiency without significantly increasing encoder and decoder complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If fixed separable transforms are used for video coding, then device complexity is reduced, but coding efficiency deteriorates due to inability to accommodate varying residual sample statistics
Solution Approach 1:
The patent implements dynamic transform selection by allowing the encoder to choose from multiple transform types (separable transforms like DCT and non-separable transforms like LFNST) based on the characteristics of residual samples. This is achieved through transform selection flags and adaptive transform indication that enable the system to switch between different transform schemes, making the transform coding process dynamic rather than fixed, thereby improving coding efficiency without significantly increasing decoder complexity.
Solution Approach 2:
The patent changes the transform parameters by introducing multiple transform candidates with different properties. The encoder can select from a set of transform types including separable transforms (DCT-2, DCT-8) and non-separable transforms (LFNST), and switch between them based on residual sample statistics. This parameter change allows the system to adapt to varying coding conditions and improve compression performance.
2Productivity
If multiple transform candidates are selected from MTS scheme, then coding efficiency is improved, but device complexity increases due to additional transform operations
Solution Approach 1:
The patent applies partial action by introducing a simplified version of LFNST that operates only on a subset of transform coefficients (specifically, only on the low-frequency coefficients in the top-left region of the transform block). This partial application of non-separable transform reduces the computational complexity compared to applying full non-separable transform, while still capturing the essential benefits of adaptive transform selection for improving coding efficiency.
Solution Approach 2:
The patent segments the transform coding process into two distinct stages: primary separable transform (applied to all coefficients) and secondary non-separable transform (LFNST applied selectively to low-frequency coefficients). This segmentation allows the system to apply complex transforms only where necessary, reducing overall computational complexity while maintaining coding efficiency benefits.
3Productivity
If secondary transforms are applied in addition to primary transforms, then compression performance is improved, but encoder and decoder complexity increases
Solution Approach 1:
The patent applies local quality by restricting the secondary LFNST operation to only the low-frequency coefficients located in the top-left region of the transform block, rather than applying it to all coefficients. This localized application reduces the number of transform operations required, thereby reducing encoder and decoder complexity while still improving compression performance where it is most needed.
Solution Approach 2:
The patent implements a nested transform structure where the secondary LFNST is applied on top of the primary separable transform coefficients. The LFNST operates on the output of the primary transform, creating a nested transformation process. This nesting allows the system to benefit from both transform types while managing complexity through selective application.
Data Source
AI summary
An example device for coding video data includes a memory configured to store video data; and one or more processors implemented in circuitry and configured to: code a first codeword representing a selected transform scheme of a set of transform candidates of a multiple transform selection (MTS) scheme for a current block of video data, the selected transform scheme being a secondary transform of a set of available secondary transforms to be applied in addition to a primary transform; code a second codeword representing the secondary transform from the set of available secondary transforms; and apply the primary transform and the secondary transform during coding of residual data for the current block. The second codeword may be a value for a low-frequency non-separable transform (LFNST) syntax element.