On debiasing restoration algorithms: applications to total-variation and nonlocal-means

TL;DR

This paper introduces a debiasing method for image restoration algorithms that reduces method bias while preserving model bias, applicable to various locally affine estimators like total variation and nonlocal filters.

Contribution

The authors propose a novel debiasing technique specifically targeting method bias in image restoration, compatible with multiple estimators.

Findings

Effective reduction of method bias demonstrated

Applicable to total variation and nonlocal filters

Preserves essential model bias for accurate restoration

Abstract

Bias in image restoration algorithms can hamper further analysis, typically when the intensities have a physical meaning of interest, e.g., in medical imaging. We propose to suppress a part of the bias -- the method bias -- while leaving unchanged the other unavoidable part -- the model bias. Our debiasing technique can be used for any locally affine estimator including \^a1 regularization, anisotropic total-variation and some nonlocal filters.