Modeling Realistic Degradations in Non-blind Deconvolution

TL;DR

This paper introduces a variational deconvolution framework that explicitly models realistic image degradations like saturation, noise, and non-linear camera response, leading to improved image restoration quality.

Contribution

It presents a novel variational approach that incorporates realistic image acquisition effects into deblurring, which was not addressed in prior methods.

Findings

Significant improvements in image quality and PSNR with realistic degradation modeling

Explicit handling of pixel saturation, noise, and gamma correction enhances deblurring accuracy

The method outperforms naive inversion approaches in restoring detailed images

Abstract

Most image deblurring methods assume an over-simplistic image formation model and as a result are sensitive to more realistic image degradations. We propose a novel variational framework, that explicitly handles pixel saturation, noise, quantization, as well as non-linear camera response function due to e.g., gamma correction. We show that accurately modeling a more realistic image acquisition pipeline leads to significant improvements, both in terms of image quality and PSNR. Furthermore, we show that incorporating the non-linear response in both the data and the regularization terms of the proposed energy leads to a more detailed restoration than a naive inversion of the non-linear curve. The minimization of the proposed energy is performed using stochastic optimization. A dataset consisting of realistically degraded images is created in order to evaluate the method.