Exploiting Epistemic Uncertainty of Anatomy Segmentation for Anomaly Detection in Retinal OCT

Abstract

Diagnosis and treatment guidance are aided by detecting relevant biomarkers in medical images. Although supervised deep learning can perform accurate segmentation of pathological areas, it is limited by requiring a-priori definitions of these regions, large-scale annotations, and a representative patient cohort in the training set. In contrast, anomaly detection is not limited to specific definitions of pathologies and allows for training on healthy samples without annotation. Anomalous regions can then serve as candidates for biomarker discovery. Knowledge about normal anatomical structure brings implicit information for detecting anomalies. We propose to take advantage of this property using bayesian deep learning, based on the assumption that epistemic uncertainties will correlate with anatomical deviations from a normal training set. A Bayesian U-Net is trained on a well-defined healthy environment using weak labels of healthy anatomy produced by existing methods. At test time, we capture epistemic uncertainty estimates of our model using Monte Carlo dropout. A novel post-processing technique is then applied to exploit these estimates and transfer their layered appearance to smooth blob-shaped segmentations of the anomalies. We experimentally validated this approach in retinal optical coherence tomography (OCT) images, using weak labels of retinal layers. Our method achieved a Dice index of 0.789 in an independent anomaly test set of age-related macular degeneration (AMD) cases. The resulting segmentations allowed very high accuracy for separating healthy and diseased cases with late wet AMD, dry geographic atrophy (GA), diabetic macular edema (DME) and retinal vein occlusion (RVO). Finally, we qualitatively observed that our approach can also detect other deviations in normal scans such as cut edge artifacts.

Accepted for publication in IEEE Transactions on Medical Imaging, 2019

References in corpus (3)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning
• Unsupervised Anomaly Detection with Generative Adversarial Networks to Guide Marker Discovery

Cited by in corpus (19)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• A Unifying Review of Deep and Shallow Anomaly Detection
• Anomaly Detection in Medical Imaging -- A Mini Review
• SegmentMeIfYouCan: A Benchmark for Anomaly Segmentation
• REB: Reducing Biases in Representation for Industrial Anomaly Detection
• Medical Imaging with Deep Learning for COVID- 19 Diagnosis: A Comprehensive Review
• The MVTec AD 2 Dataset: Advanced Scenarios for Unsupervised Anomaly Detection
• Unsupervised 3D Brain Anomaly Detection
• A Quantitative Comparison of Epistemic Uncertainty Maps Applied to Multi-Class Segmentation
• Influence of uncertainty estimation techniques on false-positive reduction in liver lesion detection
• Visual Acuity Prediction on Real-Life Patient Data Using a Machine Learning Based Multistage System
• Label-Free Segmentation of COVID-19 Lesions in Lung CT
• Constrained Contrastive Distribution Learning for Unsupervised Anomaly Detection and Localisation in Medical Images
• Analyzing Epistemic and Aleatoric Uncertainty for Drusen Segmentation in Optical Coherence Tomography Images
• Anomaly Detection using Generative Models and Sum-Product Networks in Mammography Scans
• Self-supervised Pseudo Multi-class Pre-training for Unsupervised Anomaly Detection and Segmentation in Medical Images
• Unsupervised Anomaly Detection in MR Images using Multi-Contrast Information
• Uncertainty-aware Generalized Adaptive CycleGAN
• Robustness via Uncertainty-aware Cycle Consistency