Many tasks make light work: learning to localise medical anomalies from multiple synthetic tasks
OA Location
Author(s)
Type
Chapter
Abstract
There is a growing interest in single-class modelling and out-of-distribution detection as fully supervised machine learning models cannot reliably identify classes not included in their training. The long tail of infinitely many out-of-distribution classes in real-world scenarios, e.g., for screening, triage, and quality control, means that it is often necessary to train single-class models that represent an expected feature distribution, e.g., from only strictly healthy volunteer data. Conventional supervised machine learning would require the collection of datasets that contain enough samples of all possible diseases in every imaging modality, which is not realistic. Self-supervised learning methods with synthetic anomalies are currently amongst the most promising approaches, alongside generative auto-encoders that analyse the residual reconstruction error. However, all methods suffer from a lack of structured validation, which makes calibration for deployment difficult and dataset-dependant. Our method alleviates this by making use of multiple visually-distinct synthetic anomaly learning tasks for both training and validation. This enables more robust training and generalisation. With our approach we can readily outperform state-of-the-art methods, which we demonstrate on exemplars in brain MRI and chest X-rays. Code is available at https://github.com/matt-baugh/many-tasks-make-light-work.
Editor(s)
Madabhushi, A
Greenspan, H
Mousavi, P
Salcudean, S
Duncan, J
Syeda-Mahmood, T
Taylor, R
Date Issued
2023-10-01
Citation
Medical Image Computing and Computer Assisted Intervention – MICCAI 2023, 2023, 14220, pp.162-172
ISBN
978-3-031-43906-3
Publisher
Springer Nature Switzerland AG
Start Page
162
End Page
172
Journal / Book Title
Lecture Notes in Computer Science
Volume
14220
Copyright Statement
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG.
Subjects
Computer Science
Computer Science, Artificial Intelligence
Computer Science, Theory & Methods
Life Sciences & Biomedicine
RADIOLOGY
Radiology, Nuclear Medicine & Medical Imaging
Science & Technology
SEGMENTATION
Technology
Publication Status
Published
Date Publish Online
2023-10-01