Measuring Robustness in Deep Learning Based Compressive Sensing

Abstract

Deep neural networks give state-of-the-art accuracy for reconstructing images from few and noisy measurements, a problem arising for example in accelerated magnetic resonance imaging (MRI). However, recent works have raised concerns that deep-learning-based image reconstruction methods are sensitive to perturbations and are less robust than traditional methods: Neural networks (i) may be sensitive to small, yet adversarially-selected perturbations, (ii) may perform poorly under distribution shifts, and (iii) may fail to recover small but important features in an image. In order to understand the sensitivity to such perturbations, in this work, we measure the robustness of different approaches for image reconstruction including trained and un-trained neural networks as well as traditional sparsity-based methods. We find, contrary to prior works, that both trained and un-trained methods are vulnerable to adversarial perturbations. Moreover, both trained and un-trained methods tuned for a particular dataset suffer very similarly from distribution shifts. Finally, we demonstrate that an image reconstruction method that achieves higher reconstruction quality, also performs better in terms of accurately recovering fine details. Our results indicate that the state-of-the-art deep-learning-based image reconstruction methods provide improved performance than traditional methods without compromising robustness.

References in corpus (18)

• Deep Convolutional Neural Network for Inverse Problems in Imaging
• Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift
• On instabilities of deep learning in image reconstruction - Does AI come at a cost?
• fastMRI: An Open Dataset and Benchmarks for Accelerated MRI
• Do ImageNet Classifiers Generalize to ImageNet?
• Distribution Matching Losses Can Hallucinate Features in Medical Image Translation
• Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction
• Compressed Sensing using Generative Models
• Advancing machine learning for MR image reconstruction with an open competition: Overview of the 2019 fastMRI challenge
• Deep Decoder: Concise Image Representations from Untrained Non-convolutional Networks
• Compressed Sensing with Deep Image Prior and Learned Regularization
• Solving Inverse Problems With Deep Neural Networks -- Robustness Included?
• Cold Case: The Lost MNIST Digits
• Invert to Learn to Invert
• Accuracy on the Line: On the Strong Correlation Between Out-of-Distribution and In-Distribution Generalization
• Data augmentation for deep learning based accelerated MRI reconstruction with limited data
• Intermediate Layer Optimization for Inverse Problems using Deep Generative Models
• Maximum Selection and Sorting with Adversarial Comparators and an Application to Density Estimation

Cited by in corpus (4)

• Robust Compressed Sensing MRI with Deep Generative Priors
• Validation and Generalizability of Self-Supervised Image Reconstruction Methods for Undersampled MRI
• Optimal MRI Undersampling Patterns for Ultimate Benefit of Medical Vision Tasks
• Towards Ultrafast MRI via Extreme k-Space Undersampling and Superresolution