Know Your Space: Inlier and Outlier Construction for Calibrating Medical OOD Detectors
This addresses the challenge of safe deployment of medical AI by improving OOD detection calibration, though it is incremental as it builds on existing energy-based frameworks and augmentation techniques.
The paper tackled the problem of calibrating out-of-distribution (OOD) detectors for medical image classifiers by finding that the synthesis space (latent vs. pixel) is critical, leading to a 15%-35% improvement in AUROC over state-of-the-art methods across benchmarks.
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.