Uncertainty Quantification in Deep Learning for Safer Neuroimage Enhancement

Abstract

Deep learning (DL) has shown great potential in medical image enhancement problems, such as super-resolution or image synthesis. However, to date, little consideration has been given to uncertainty quantification over the output image. Here we introduce methods to characterise different components of uncertainty in such problems and demonstrate the ideas using diffusion MRI super-resolution. Specifically, we propose to account for $intrinsic$ uncertainty through a heteroscedastic noise model and for $parameter$ uncertainty through approximate Bayesian inference, and integrate the two to quantify $predictive$ uncertainty over the output image. Moreover, we introduce a method to propagate the predictive uncertainty on a multi-channelled image to derived scalar parameters, and separately quantify the effects of intrinsic and parameter uncertainty therein. The methods are evaluated for super-resolution of two different signal representations of diffusion MR images---DTIs and Mean Apparent Propagator MRI---and their derived quantities such as MD and FA, on multiple datasets of both healthy and pathological human brains. Results highlight three key benefits of uncertainty modelling for improving the safety of DL-based image enhancement systems. Firstly, incorporating uncertainty improves the predictive performance even when test data departs from training data. Secondly, the predictive uncertainty highly correlates with errors, and is therefore capable of detecting predictive "failures". Results demonstrate that such an uncertainty measure enables subject-specific and voxel-wise risk assessment of the output images. Thirdly, we show that the method for decomposing predictive uncertainty into its independent sources provides high-level "explanations" for the performance by quantifying how much uncertainty arises from the inherent difficulty of the task or the limited training examples.

References in corpus (24)

• A Survey on Deep Learning in Medical Image Analysis
• Auto-Encoding Variational Bayes
• Overcoming catastrophic forgetting in neural networks
• Efficient Multi-Scale 3D CNN with Fully Connected CRF for Accurate Brain Lesion Segmentation
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• Conditional Image Synthesis With Auxiliary Classifier GANs
• Image reconstruction by domain transform manifold learning
• Variational Inference with Normalizing Flows
• Anatomically Constrained Neural Networks (ACNN): Application to Cardiac Image Enhancement and Segmentation
• Stochastic Gradient Hamiltonian Monte Carlo
• Variational Dropout and the Local Reparameterization Trick
• Unsupervised Learning for Fast Probabilistic Diffeomorphic Registration
• Variational Dropout Sparsifies Deep Neural Networks
• Distribution Matching Losses Can Hallucinate Features in Medical Image Translation
• MR fingerprinting Deep RecOnstruction NEtwork (DRONE)
• Multiplicative Normalizing Flows for Variational Bayesian Neural Networks
• Cyclical Stochastic Gradient MCMC for Bayesian Deep Learning
• Structured and Efficient Variational Deep Learning with Matrix Gaussian Posteriors
• Cost-Effective Active Learning for Melanoma Segmentation
• Adversarial training with cycle consistency for unsupervised super-resolution in endomicroscopy
• Implicit Weight Uncertainty in Neural Networks
• Towards continual learning in medical imaging
• Radial and Directional Posteriors for Bayesian Neural Networks

Cited by in corpus (4)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• A Decoupled Uncertainty Model for MRI Segmentation Quality Estimation
• Sampling possible reconstructions of undersampled acquisitions in MR imaging
• SUREMap: Predicting Uncertainty in CNN-based Image Reconstruction Using Stein's Unbiased Risk Estimate