Blind Image Restoration with Flow Based Priors

TL;DR

This paper introduces a novel blind image restoration method using normalizing flows as priors within a MAP framework, enabling effective handling of unknown degradations with competitive results.

Contribution

It is the first to utilize normalizing flows as priors for image enhancement, expressing MAP optimization in latent space for improved blind restoration.

Findings

Achieved competitive results against deep image prior methods.

Demonstrated effectiveness across various degradation types.

Validated on datasets of varying complexity.

Abstract

Image restoration has seen great progress in the last years thanks to the advances in deep neural networks. Most of these existing techniques are trained using full supervision with suitable image pairs to tackle a specific degradation. However, in a blind setting with unknown degradations this is not possible and a good prior remains crucial. Recently, neural network based approaches have been proposed to model such priors by leveraging either denoising autoencoders or the implicit regularization captured by the neural network structure itself. In contrast to this, we propose using normalizing flows to model the distribution of the target content and to use this as a prior in a maximum a posteriori (MAP) formulation. By expressing the MAP optimization process in the latent space through the learned bijective mapping, we are able to obtain solutions through gradient descent. To the best of our knowledge, this is the first work that explores normalizing flows as prior in image enhancement problems. Furthermore, we present experimental results for a number of different degradations on data sets varying in complexity and show competitive results when comparing with the deep image prior approach.