Inverse Quantization Matrix Zeroing for Video Decoder Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increased size of transform blocks in new video coding standards, such as 64×64, complicates the inverse transform process in video coding/decoding, making decoder implementation more difficult due to higher computational complexity and excessive inverse transform coefficients.
Innovation Solution
A method for inverse quantization that involves determining an initial weighted inverse quantization matrix, setting specific matrix elements to zero based on the size of the quantized block, and using these weights to generate inverse transform coefficients, thereby controlling the distribution of non-zero coefficients in the inverse transform block and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the size of the transform block is increased to support 4K TV technology and new video coding standards, then the video coding capability is improved, but the complexity of the inverse transform increases and decoder implementation becomes more difficult
Solution Approach 1:
The patent extracts and eliminates unnecessary inverse transform coefficients by setting matrix elements to zero based on threshold conditions. Specifically, elements M[x][y] are set to zero when x >= Tx or y >= Ty, removing redundant coefficients that increase computational complexity while maintaining essential video coding functionality.
Solution Approach 2:
The patent applies local quality by differentiating the treatment of different matrix elements based on their positions. Different threshold values Tx and Ty are applied to different regions of the matrix, creating a non-uniform pattern of zeroing that adapts to the specific characteristics of each block region, thereby optimizing the balance between coding capability and computational complexity.
2Adaptability or versatility
If the size of the transform block is increased to 64×64, then the video coding standard capability is improved, but the number of inverse transform coefficients increases excessively
Solution Approach 1:
The patent removes excessive coefficients by applying threshold-based zeroing to the inverse quantization matrix. For a 64×64 transform block, the method sets elements to zero when their row or column indices exceed predetermined thresholds, significantly reducing the number of non-zero coefficients while preserving the essential information needed for video decoding.
Solution Approach 2:
The patent applies partial action by selectively zeroing only the portions of the matrix that exceed the thresholds Tx and Ty, rather than zeroing the entire matrix or using uniform thresholding. This partial zeroing approach removes excessive coefficients while preserving the necessary coefficients for maintaining video quality.
3Device complexity
If the inverse transform block size is reduced by zeroing matrix elements, then the computational complexity is reduced, but the precision of the inverse transform may be affected
Solution Approach 1:
The patent changes the parameters of the inverse quantization matrix by introducing threshold values Tx and Ty that determine which elements are zeroed. By carefully selecting these threshold parameters, the method reduces computational complexity while maintaining sufficient precision for video decoding, as the thresholds are designed to preserve the most significant coefficients.
Solution Approach 2:
The patent applies local quality by differentiating the importance of different matrix elements through position-dependent thresholding. Elements in certain regions are zeroed while others are preserved, based on their local significance to the transform. This localized approach maintains precision in critical regions while reducing complexity in less important regions.
Data Source
AI summary
A method, system, device and computer-readable storage medium for inverse quantization. The method comprises: determining an initial weighted inverse quantization matrix, wherein, the initial weighted inverse quantization matrix is the same as the quantized block in size; setting some matrix elements in the initial weighted inverse quantization matrix to zero to obtain a weighted inverse quantization matrix, wherein, determining the matrix elements that need to be zeroed according to the size of the quantized block; weighted inverse quantizing the quantized coefficients in the quantized block to generate corresponding inverse transform coefficients, wherein, the value of the matrix element corresponding to the position of the quantized coefficient in the weighted inverse quantization matrix is used as a weight coefficient of the weighted inverse quantization.