Semi-sparsity Priors for Image Structure Analysis and Extraction

TL;DR

This paper introduces a semi-sparse regularization framework for image structure-texture decomposition, effectively separating structures from textures, avoiding artifacts, and providing a versatile, efficient solution with superior results.

Contribution

The paper proposes a novel semi-sparse regularization method for image decomposition that improves structure preservation and texture separation, with an efficient ADMM-based numerical solution.

Findings

Effective preservation of image structures without staircase artifacts

Capability to decompose complex oscillatory textures

Achieves comparable or superior results to state-of-the-art methods

Abstract

Image structure-texture decomposition is a long-standing and fundamental problem in both image processing and computer vision fields. In this paper, we propose a generalized semi-sparse regularization framework for image structural analysis and extraction, which allows us to decouple the underlying image structures from complicated textural backgrounds. Combining with different textural analysis models, such a regularization receives favorable properties differing from many traditional methods. We demonstrate that it is not only capable of preserving image structures without introducing notorious staircase artifacts in polynomial-smoothing surfaces but is also applicable for decomposing image textures with strong oscillatory patterns. Moreover, we also introduce an efficient numerical solution based on an alternating direction method of multipliers (ADMM) algorithm, which gives rise to a simple and maneuverable way for image structure-texture decomposition. The versatility of the proposed method is finally verified by a series of experimental results with the capability of producing comparable or superior image decomposition results against cutting-edge methods.