DeepRED: Deep Image Prior Powered by RED

Abstract

Inverse problems in imaging are extensively studied, with a variety of strategies, tools, and theory that have been accumulated over the years. Recently, this field has been immensely influenced by the emergence of deep-learning techniques. One such contribution, which is the focus of this paper, is the Deep Image Prior (DIP) work by Ulyanov, Vedaldi, and Lempitsky (2018). DIP offers a new approach towards the regularization of inverse problems, obtained by forcing the recovered image to be synthesized from a given deep architecture. While DIP has been shown to be quite an effective unsupervised approach, its results still fall short when compared to state-of-the-art alternatives. In this work, we aim to boost DIP by adding an explicit prior, which enriches the overall regularization effect in order to lead to better-recovered images. More specifically, we propose to bring-in the concept of Regularization by Denoising (RED), which leverages existing denoisers for regularizing inverse problems. Our work shows how the two (DIP and RED) can be merged into a highly effective unsupervised recovery process while avoiding the need to differentiate the chosen denoiser, and leading to very effective results, demonstrated for several tested problems.

References in corpus (5)

• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Fast Image Recovery Using Variable Splitting and Constrained Optimization
• Deep Image Prior
• Sparsity Based Poisson Denoising with Dictionary Learning
• Unsupervised Learning with Stein's Unbiased Risk Estimator

Cited by in corpus (12)

• RARE: Image Reconstruction using Deep Priors Learned without Ground Truth
• Learning to Optimize: A Primer and A Benchmark
• SIMBA: Scalable Inversion in Optical Tomography using Deep Denoising Priors
• $\Sigma$-net: Systematic Evaluation of Iterative Deep Neural Networks for Fast Parallel MR Image Reconstruction
• Solving Linear Inverse Problems Using the Prior Implicit in a Denoiser
• Async-RED: A Provably Convergent Asynchronous Block Parallel Stochastic Method using Deep Denoising Priors
• Joint Reconstruction and Calibration using Regularization by Denoising
• Towards the Automation of Deep Image Prior
• Rethinking Deep Image Prior for Denoising
• Across-domains transferability of Deep-RED in de-noising and compressive sensing recovery of seismic data
• Untrained Graph Neural Networks for Denoising
• Blind Motion Deblurring through SinGAN Architecture