Deep Unfolding with Normalizing Flow Priors for Inverse Problems

Abstract

Many application domains, spanning from computational photography to medical imaging, require recovery of high-fidelity images from noisy, incomplete or partial/compressed measurements. State of the art methods for solving these inverse problems combine deep learning with iterative model-based solvers, a concept known as deep algorithm unfolding. By combining a-priori knowledge of the forward measurement model with learned (proximal) mappings based on deep networks, these methods yield solutions that are both physically feasible (data-consistent) and perceptually plausible. However, current proximal mappings only implicitly learn such image priors. In this paper, we propose to make these image priors fully explicit by embedding deep generative models in the form of normalizing flows within the unfolded proximal gradient algorithm. We demonstrate that the proposed method outperforms competitive baselines on various image recovery tasks, spanning from image denoising to inpainting and deblurring.

References in corpus (11)

• Auto-Encoding Variational Bayes
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• Normalizing Flows: An Introduction and Review of Current Methods
• Learned Primal-dual Reconstruction
• EdgeConnect: Generative Image Inpainting with Adversarial Edge Learning
• Denoising Prior Driven Deep Neural Network for Image Restoration
• Deep Unfolding: Model-Based Inspiration of Novel Deep Architectures
• Learned D-AMP: Principled Neural Network based Compressive Image Recovery
• Unrolled Optimization with Deep Priors
• Neural Proximal Gradient Descent for Compressive Imaging
• Invert to Learn to Invert

Cited by in corpus (3)

• PatchNR: Learning from Very Few Images by Patch Normalizing Flow Regularization
• Top-philic Machine Learning
• Monte Carlo guided Diffusion for Bayesian linear inverse problems