Neumann Networks for Inverse Problems in Imaging

Abstract

Many challenging image processing tasks can be described by an ill-posed linear inverse problem: deblurring, deconvolution, inpainting, compressed sensing, and superresolution all lie in this framework. Traditional inverse problem solvers minimize a cost function consisting of a data-fit term, which measures how well an image matches the observations, and a regularizer, which reflects prior knowledge and promotes images with desirable properties like smoothness. Recent advances in machine learning and image processing have illustrated that it is often possible to learn a regularizer from training data that can outperform more traditional regularizers. We present an end-to-end, data-driven method of solving inverse problems inspired by the Neumann series, which we call a Neumann network. Rather than unroll an iterative optimization algorithm, we truncate a Neumann series which directly solves the linear inverse problem with a data-driven nonlinear regularizer. The Neumann network architecture outperforms traditional inverse problem solution methods, model-free deep learning approaches, and state-of-the-art unrolled iterative methods on standard datasets. Finally, when the images belong to a union of subspaces and under appropriate assumptions on the forward model, we prove there exists a Neumann network configuration that well-approximates the optimal oracle estimator for the inverse problem and demonstrate empirically that the trained Neumann network has the form predicted by theory.

Added further experiments, reorganized proof section, added further references and supporting figures

References in corpus (14)

• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Robust Recovery of Subspace Structures by Low-Rank Representation
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Low-Dose CT with a Residual Encoder-Decoder Convolutional Neural Network (RED-CNN)
• Trainable Nonlinear Reaction Diffusion: A Flexible Framework for Fast and Effective Image Restoration
• Compressed Sensing using Generative Models
• Convergence Rates of Inexact Proximal-Gradient Methods for Convex Optimization
• Learned D-AMP: Principled Neural Network based Compressive Image Recovery
• Unrolled Optimization with Deep Priors
• Neural Proximal Gradient Descent for Compressive Imaging
• Deep BCD-Net Using Identical Encoding-Decoding CNN Structures for Iterative Image Recovery
• Conditional density estimation in a regression setting
• MoDL-MUSSELS: Model-Based Deep Learning for Multi-Shot Sensitivity Encoded Diffusion MRI
• Reconstruction of Signals Drawn from a Gaussian Mixture from Noisy Compressive Measurements

Cited by in corpus (7)

• Solving Inverse Problems by Joint Posterior Maximization with Autoencoding Prior
• Bayesian imaging using Plug & Play priors: when Langevin meets Tweedie
• Solving Inverse Problems by Joint Posterior Maximization with a VAE Prior
• Solving the Wide-band Inverse Scattering Problem via Equivariant Neural Networks
• MRI Image Reconstruction via Learning Optimization Using Neural ODEs
• Deep Unrolled Network for Video Super-Resolution
• AHP-Net: adaptive-hyper-parameter deep learning based image reconstruction method for multilevel low-dose CT