Non-Separable Transform Kernel Selection for Large Video Blocks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding techniques face challenges in achieving higher coding efficiency and improved image enhancement due to increasing image size, resolution, and frame rate, particularly with the non-separable primary transform (NSPT) requiring improvements based on intra prediction mode, transform block size, and transform coefficient characteristics.
Innovation Solution
A video coding method and apparatus that perform non-separable primary transform (NSPT) based on the intra prediction mode, transform block size, and transform coefficient characteristics, using implicit partitioning for large blocks, and applying non-separable primary inverse transforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If non-separable primary transform is applied to large transform blocks, then transform efficiency is improved, but the transform cannot be directly applied due to size constraints
Solution Approach 1:
The patent applies segmentation by dividing large transform blocks into multiple sub-blocks that can be processed individually using non-separable primary transform. The transform block is partitioned into smaller regions, allowing the transform to be applied to each sub-block separately, thus overcoming the size constraint while maintaining transform efficiency.
2Measurement precision
If more transform kernels are defined for different block shapes and intra prediction modes, then transform precision is improved, but system complexity increases
Solution Approach 1:
The patent applies local quality by selecting different transform kernels based on local characteristics such as block shape and intra prediction mode. Instead of using a single transform kernel for all cases, the system adapts the transform kernel to match the local properties of each transform block, improving transform precision while managing complexity through conditional selection rather than maintaining all possible kernels simultaneously.
3Manufacturing precision
If image size, resolution, and frame rate are increased, then video quality is improved, but data amount to be encoded increases
Solution Approach 1:
The patent applies parameter changes by modifying the transform process parameters (transform kernel selection, transform block size, non-separable transform application) based on the characteristics of the video content. This allows the encoding system to adapt to higher resolution and frame rate requirements by optimizing transform parameters for different block sizes and prediction modes, improving compression efficiency and reducing the overall data amount while maintaining high video quality.
Data Source
AI summary
A method and an apparatus are disclosed for video coding based on non-separable primary transform. In the disclosed embodiments, a video decoding device acquires inversely-quantized transform coefficients for a transform block of the current block and decodes a Non-separable Primary Transform (NSPT) flag from a bitstream. The video decoding device checks the NSPT flag and, when the NSPT flag is true, determines a non-separable primary inverse transform kernel based on a size of the transform block, an intra prediction mode of the current block, and characteristics of the inversely-quantized transform coefficients. The video decoding device generates residual signals by applying the non-separable primary inverse transform kernel to the transform coefficients.