Equivariant Imaging: Learning Beyond the Range Space

Abstract

In various imaging problems, we only have access to compressed measurements of the underlying signals, hindering most learning-based strategies which usually require pairs of signals and associated measurements for training. Learning only from compressed measurements is impossible in general, as the compressed observations do not contain information outside the range of the forward sensing operator. We propose a new end-to-end self-supervised framework that overcomes this limitation by exploiting the equivariances present in natural signals. Our proposed learning strategy performs as well as fully supervised methods. Experiments demonstrate the potential of this framework on inverse problems including sparse-view X-ray computed tomography on real clinical data and image inpainting on natural images. Code has been made available at: https://github.com/edongdongchen/EI.

arXiv commentICCV 2021. Code: https://github.com/edongdongchen/EI

References in corpus (9)

• Adam: A Method for Stochastic Optimization
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• Low Dose CT Image Denoising Using a Generative Adversarial Network with Wasserstein Distance and Perceptual Loss
• Image Restoration Using Very Deep Convolutional Encoder-Decoder Networks with Symmetric Skip Connections
• ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks
• RARE: Image Reconstruction using Deep Priors Learned without Ground Truth
• Neural Proximal Gradient Descent for Compressive Imaging
• Generalization Error of Invariant Classifiers
• Improving Transformation Invariance in Contrastive Representation Learning