Interpretable CNN pruning for preserving scale-covariant features in medical imaging

Graziani, Mara (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis) ; University of Geneva, Switzerland) ; Lompech, Thomas (INP-ENSEEIHT, Toulouse, France) ; Müller, Henning (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis) ; University of Geneva, Switzerland) ; Depeursinge, Adrien (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis) ; Centre Hospitalier Universitaire Vaudois (CHUV), Lausanne, Switzerland) ; Andrearczyk, Vincent (University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis))

Image scale carries crucial information in medical imaging, e.g. the size and spatial frequency of local structures, lesions, tumors and cell nuclei. With feature transfer being a common practice, scale-invariant features implicitly learned from pretraining on ImageNet tend to be preferred over scale-covariant features. The pruning strategy in this paper proposes a way to maintain scale covariance in the transferred features. Deep learning interpretability is used to analyze the layer-wise encoding of scale information for popular architectures such as InceptionV3 and ResNet50. Interestingly, the covariance of scale peaks at central layers and decreases close to softmax. Motivated by these results, our pruning strategy removes the layers where invariance to scale is learned. The pruning operation leads to marked improvements in the regression of both nuclei areas and magnification levels of histopathology images. These are relevant applications to enlarge the existing medical datasets with open-access images as those of PubMed Central. All experiments are performed on publicly available data and the code is shared on GitHub.


Note: Due to the COVID-19 outbreak, the Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2020 venue in Lima was cancelled. The proceedings of the online conference are however published according to the original schedule.


Keywords:
Conference Type:
published full paper
Faculty:
Economie et Services
School:
HEG-VS
Institute:
Institut Informatique de gestion
Publisher:
Lima, Peru, 4-8 October 2020
Date:
2020-10
Lima, Peru
4-8 October 2020
Pagination:
Pp. 23-32
Published in:
Proceedings of Workshop on Interpretability of Machine Intelligence in Medical Image Computing at MICCAI 2020
Series Statement:
Lecture Notes in Computer Science, vol. 12446
DOI:
ISSN:
0302-9743
ISBN:
978-3-030-61165-1
Appears in Collection:

Note: The status of this file is: restricted


 Record created 2021-01-26, last modified 2021-01-29

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)