Well-Calibrated Regression Uncertainty in Medical Imaging with Deep Learning

verfasst von

Max Heinrich Laves, Sontje Ihler, Jacob F. Fast, Lüder A. Kahrs, Tobias Ortmaier

Abstract

The consideration of predictive uncertainty in medical imaging with deep learning is of utmost importance. We apply estimation of predictive uncertainty by variational Bayesian inference with Monte Carlo dropout to regression tasks and show why predictive uncertainty is systematically underestimated. We suggest using σ scaling with a single scalar value; a simple, yet effective calibration method for both aleatoric and epistemic uncertainty. The performance of our approach is evaluated on a variety of common medical regression data sets using different state-of-the-art convolutional network architectures. In all experiments, σ scaling is able to reliably recalibrate predictive uncertainty. It is easy to implement and maintains the accuracy. Well-calibrated uncertainty in regression allows robust rejection of unreliable predictions or detection of out-of-distribution samples. Our source code is available at: github.com/mlaves/well-calibrated-regression-uncertainty.

Organisationseinheit(en)

Institut für Mechatronische Systeme

Externe Organisation(en)

Centre for Image Guided Innovation and Therapeutic Intervention (CIGITI)
University of Toronto

Typ

Konferenzaufsatz in Fachzeitschrift

Journal

Proceedings of Machine Learning Research

Band

121

Seiten

393-412

Anzahl der Seiten

20

Publikationsdatum

2020

Publikationsstatus

Veröffentlicht

Peer-reviewed

Ja

ASJC Scopus Sachgebiete

Artificial intelligence, Software, Steuerungs- und Systemtechnik, Statistik und Wahrscheinlichkeit

Elektronische Version(en)

https://proceedings.mlr.press/v121/laves20a.html (Zugang: Offen)