Share Email Print

Proceedings Paper

Supervised novelty detection in brain tissue classification with an application to white matter hyperintensities
Author(s): Hugo J. Kuijf; Pim Moeskops; Bob D. de Vos; Willem H. Bouvy; Jeroen de Bresser; Geert Jan Biessels; Max A. Viergever; Koen L. Vincken
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Novelty detection is concerned with identifying test data that differs from the training data of a classifier. In the case of brain MR images, pathology or imaging artefacts are examples of untrained data. In this proof-of-principle study, we measure the behaviour of a classifier during the classification of trained labels (i.e. normal brain tissue). Next, we devise a measure that distinguishes normal classifier behaviour from abnormal behavior that occurs in the case of a novelty. This will be evaluated by training a kNN classifier on normal brain tissue, applying it to images with an untrained pathology (white matter hyperintensities (WMH)), and determine if our measure is able to identify abnormal classifier behaviour at WMH locations. For our kNN classifier, behaviour is modelled as the mean, median, or q1 distance to the k nearest points. Healthy tissue was trained on 15 images; classifier behaviour was trained/tested on 5 images with leave-one-out cross-validation. For each trained class, we measure the distribution of mean/median/q1 distances to the k nearest point. Next, for each test voxel, we compute its Z-score with respect to the measured distribution of its predicted label. We consider a Z-score ≥4 abnormal behaviour of the classifier, having a probability due to chance of 0.000032. Our measure identified >90% of WMH volume and also highlighted other non-trained findings. The latter being predominantly vessels, cerebral falx, brain mask errors, choroid plexus. This measure is generalizable to other classifiers and might help in detecting unexpected findings or novelties by measuring classifier behaviour.

Paper Details

Date Published: 21 March 2016
PDF: 7 pages
Proc. SPIE 9784, Medical Imaging 2016: Image Processing, 978421 (21 March 2016); doi: 10.1117/12.2216358
Show Author Affiliations
Hugo J. Kuijf, Univ. Medical Ctr. Utrecht (Netherlands)
Pim Moeskops, Univ. Medical Ctr. Utrecht (Netherlands)
Bob D. de Vos, Univ. Medical Ctr. Utrecht (Netherlands)
Willem H. Bouvy, Univ. Medical Ctr. Utrecht (Netherlands)
Jeroen de Bresser, Univ. Medical Ctr. Utrecht (Netherlands)
Geert Jan Biessels, Univ. Medical Ctr. Utrecht (Netherlands)
Max A. Viergever, Univ. Medical Ctr. Utrecht (Netherlands)
Koen L. Vincken, Univ. Medical Ctr. Utrecht (Netherlands)

Published in SPIE Proceedings Vol. 9784:
Medical Imaging 2016: Image Processing
Martin A. Styner; Elsa D. Angelini, Editor(s)

© SPIE. Terms of Use
Back to Top