Non-Separable Transform Matrix for Video Signal Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding technologies face challenges in efficiently processing next-generation video content with high spatial resolution, high frame rate, and high dimensionality, requiring an efficient transform technique for spatial to frequency domain conversion with low complexity.
Innovation Solution
A method and apparatus for decoding image signals using a non-separable transform matrix, determining input and output lengths based on block sizes, and applying this matrix to coefficients, with specific configurations for block sizes greater than or equal to 8, including adaptive transform sets and kernels for intra prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional transform techniques are used for next-generation video content, then processing capability is sufficient for current standards, but processing power and complexity will be insufficient for high spatial resolution and high frame rate content
Solution Approach 1:
The transform process is segmented into multiple stages: primary transform, secondary transform, and inverse transforms. Each stage processes specific frequency components separately, allowing independent optimization of each transform operation to reduce overall computational complexity while maintaining processing capability for high resolution content
Solution Approach 2:
The patent implements dynamic transform selection based on block size and prediction mode. Different transform types (DST, DCT, DIT) are dynamically selected according to the characteristics of the current block, enabling adaptive processing that optimizes both processing power utilization and computational complexity for varying content types
2Productivity
If transform size is increased to handle high spatial resolution, then processing capability improves, but memory storage and access requirements increase
Solution Approach 1:
Large transform blocks are segmented into smaller sub-blocks for processing. The primary transform is applied to the entire block, then the secondary transform is applied to sub-blocks, reducing the memory footprint required for any single transform operation while maintaining the ability to process high spatial resolution content
Solution Approach 2:
The patent applies transforms in multiple dimensions sequentially (horizontal and vertical directions). This dimensional decomposition allows processing of large high-resolution blocks by breaking them down into smaller 2D transform operations, reducing peak memory storage requirements
3Manufacturing precision
If transform accuracy is increased for better compression, then coding efficiency improves, but processing complexity increases
Solution Approach 1:
The patent dynamically selects transform types and sizes based on block characteristics and prediction modes. This adaptive approach achieves high coding efficiency by using the most appropriate transform for each block while avoiding the complexity of applying high-precision transforms uniformly across all blocks
Solution Approach 2:
The patent changes transform parameters (type, size, direction) based on content characteristics. By adjusting these parameters adaptively rather than using fixed high-precision transforms, the system achieves high coding efficiency while controlling processing complexity through selective application
Data Source
AI summary
The embodiments of the present disclosure provides a method and apparatus for video signal processing. A method for decoding an image signal according to an embodiment of the present disclosure may include determining an input length and an output length of a non-separable transform based on a height and a width of a current block; determining a non-separable transform matrix corresponding to the input length and the output length of a non-separable transform; and applying the non-separable transform matrix to coefficients by a number of the input length in the current block, wherein the height and the width of a current block is greater than or equal to 8, wherein, if each of the height and the width of a current block is equal to 8, the input length of the non-separable transform is determined as 8.