Abstract
We focus on the problem of producing well-calibrated out-of-distribution (OOD) detectors, in order to enable safe deployment of medical image classifiers. Motivated by the difficulty of curating suitable calibration datasets, synthetic augmentations have become highly prevalent for inlier/outlier specification. While there have been rapid advances in data augmentation techniques, this paper makes a striking finding that the space in which the inliers and outliers are synthesized, in addition to the type of augmentation, plays a critical role in calibrating OOD detectors. Using the popular energy-based OOD detection framework, we find that the optimal protocol is to synthesize latent-space inliers along with diverse pixel-space outliers. Based on empirical studies with multiple medical imaging benchmarks, we demonstrate that our approach consistently leads to superior OOD detection (15% − 35% in AUROC) over the state-of-the-art in a variety of open-set recognition settings.
Original language | English (US) |
---|---|
Pages (from-to) | 190-211 |
Number of pages | 22 |
Journal | Proceedings of Machine Learning Research |
Volume | 227 |
State | Published - 2023 |
Event | 6th International Conference on Medical Imaging with Deep Learning, MIDL 2023 - Nashville, United States Duration: Jul 10 2023 → Jul 12 2023 |
Keywords
- data augmentation
- Deep neural networks
- energy
- medical imaging
- open-set recognition
- out-of-distribution detection
ASJC Scopus subject areas
- Artificial Intelligence
- Software
- Control and Systems Engineering
- Statistics and Probability