Regularizing linear inverse problems with convolutional neural networks

Abstract

Deep convolutional neural networks trained on large datsets have emerged as an intriguing alternative for compressing images and solving inverse problems such as denoising and compressive sensing. However, it has only recently been realized that even without training, convolutional networks can function as concise image models, and thus regularize inverse problems. In this paper, we provide further evidence for this finding by studying variations of convolutional neural networks that map few weight parameters to an image. The networks we consider only consist of convolutional operations, with either fixed or parameterized filters followed by ReLU non-linearities. We demonstrate that with both fixed and parameterized convolutional filters those networks enable representing images with few coefficients. What is more, the underparameterization enables regularization of inverse problems, in particular recovering an image from few observations. We show that, similar to standard compressive sensing guarantees, on the order of the number of model parameters many measurements suffice for recovering an image from compressive measurements. Finally, we demonstrate that signal recovery with a un-trained convolutional network outperforms standard l1 and total variation minimization for magnetic resonance imaging (MRI).

References in corpus (7)

• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• fastMRI: An Open Dataset and Benchmarks for Accelerated MRI
• Variational Autoencoder for Deep Learning of Images, Labels and Captions
• Lossy Image Compression with Compressive Autoencoders
• Compressed Sensing with Deep Image Prior and Learned Regularization
• Algorithmic Guarantees for Inverse Imaging with Untrained Network Priors
• Channel Normalization in Convolutional Neural Network avoids Vanishing Gradients

Cited by in corpus (5)

• Algorithmic Guarantees for Inverse Imaging with Untrained Network Priors
• Denoising and Regularization via Exploiting the Structural Bias of Convolutional Generators
• Generative Models for Low-Rank Video Representation and Reconstruction
• Compressive sensing with un-trained neural networks: Gradient descent finds the smoothest approximation
• Subsampled Fourier Ptychography using Pretrained Invertible and Untrained Network Priors