Separable Adaptive Loop Filter for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing adaptive loop filtering methods for image and video signals require significant computational overhead in decoders, compromising perceptual quality due to reliance on two-dimensional Wiener filters.
Innovation Solution
Implementing an optimal separable Wiener filter as a good approximation of non-separable filters, reducing decoder computational complexity while maintaining video quality by using a product of one-dimensional vertical and horizontal filters, which are iteratively improved to minimize sum of squares error.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If two-dimensional Wiener filters are used for adaptive loop filtering, then filtering performance is improved, but decoder computational complexity increases significantly
Solution Approach 1:
The patent divides the two-dimensional Wiener filter into separable components by expressing it as a product of one-dimensional filters. Specifically, the 2D filter coefficients are decomposed into horizontal and vertical 1D filter coefficients that can be applied sequentially, reducing the computational complexity from O(N²) to O(N) operations while maintaining the filtering performance.
Solution Approach 2:
The patent transforms the filter representation from a non-separable 2D kernel to a separable form with different parameter organization. By changing how the filter coefficients are structured and computed (from a single 2D array to separate 1D arrays), the system achieves the same filtering effect with reduced computational burden.
2Device complexity
If separable filters are used to reduce computational complexity, then decoder processing is simplified, but filter accuracy may be compromised
Solution Approach 1:
The patent performs preliminary computation of optimal separable filter coefficients during the encoding stage. The encoder calculates the separable filter parameters based on the block type and surrounding block characteristics, then transmits these pre-computed parameters to the decoder, eliminating the need for complex real-time computations at decoding time while maintaining high filter accuracy.
Solution Approach 2:
The patent creates a simplified copy of the 2D Wiener filter in separable form that approximates the original non-separable filter's behavior. By carefully constructing the separable filter coefficients to match the 2D filter's frequency response and filtering characteristics, the system achieves accurate approximation with reduced complexity.
Data Source
AI summary
An adaptive loop filter utilizing separable filters within a quadtree-based adaptive loop filter (QALF). The filters for at least the P and B pictures are replaced with a separable filter to provide Separable QALF (SQALF). In a preferred embodiment the filter comprises a Wiener filter as a best fit approximation of the non-separable filter. In response to decoding with the SQALF approach, computational complexity in the decoder is reduced without lowering perceived visual quality. Estimation of the separable filter is obtained in the encoder by iteratively minimizing the sum of squares error with the separable filter. The SQALF filtering is described within a video decoder.