Nonlocal Adaptive Direction-Guided Structure Tensor Total Variation For Image Recovery

TL;DR

This paper enhances nonlocal structure tensor total variation (NLSTV) for image recovery by incorporating directional priors estimated via anisotropic Gaussian kernels, leading to improved visual and quantitative results.

Contribution

It introduces a novel two-stage framework that boosts NLSTV regularization with directional information derived from anisotropic Gaussian kernels, improving image recovery performance.

Findings

Outperforms NLSTV and other local models in experiments

Achieves better visual quality in recovered images

Provides quantitative improvements over existing methods

Abstract

A common strategy in variational image recovery is utilizing the nonlocal self-similarity (NSS) property, when designing energy functionals. One such contribution is nonlocal structure tensor total variation (NLSTV), which lies at the core of this study. This paper is concerned with boosting the NLSTV regularization term through the use of directional priors. More specifically, NLSTV is leveraged so that, at each image point, it gains more sensitivity in the direction that is presumed to have the minimum local variation. The actual difficulty here is capturing this directional information from the corrupted image. In this regard, we propose a method that employs anisotropic Gaussian kernels to estimate directional features to be later used by our proposed model. The experiments validate that our entire two-stage framework achieves better results than the NLSTV model and two other competing local models, in terms of visual and quantitative evaluation.