Efficient Out-of-Distribution Detection in Digital Pathology Using Multi-Head Convolutional Neural Networks

Jasper Linmans, Jeroen van der Laak, Geert Litjens
Proceedings of the Third Conference on Medical Imaging with Deep Learning, PMLR 121:465-478, 2020.

Abstract

Successful clinical implementation of deep learning in medical imaging depends, in part, on the reliability of the predictions. Specifically, the system should be accurate for classes seen during training while providing calibrated estimates of uncertainty for abnormalities and unseen classes. To efficiently estimate predictive uncertainty, we propose the use of multi-head CNNs (M-heads). We compare its performance to related and more prevalent approaches, such as deep ensembles, on the task of out-of-distribution (OOD) detection. To this end, we evaluate models trained to discriminate normal lymph node tissue from breast cancer metastases, on lymph nodes containing lymphoma. We show the ability to discriminate between in-distribution lymph node tissue and lymphoma by evaluating the AUROC based on the uncertainty signal. Here, the best performing multi-head CNN (91.7) outperforms both Monte Carlo dropout (88.3) and deep ensembles (86.8). Furthermore, we show that the meta-loss function of M-heads improves OOD detection in terms of AUROC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v121-linmans20a, title = {Efficient Out-of-Distribution Detection in Digital Pathology Using Multi-Head Convolutional Neural Networks}, author = {Linmans, Jasper and van der Laak, Jeroen and Litjens, Geert}, booktitle = {Proceedings of the Third Conference on Medical Imaging with Deep Learning}, pages = {465--478}, year = {2020}, editor = {Arbel, Tal and Ben Ayed, Ismail and de Bruijne, Marleen and Descoteaux, Maxime and Lombaert, Herve and Pal, Christopher}, volume = {121}, series = {Proceedings of Machine Learning Research}, month = {06--08 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v121/linmans20a/linmans20a.pdf}, url = {https://proceedings.mlr.press/v121/linmans20a.html}, abstract = {Successful clinical implementation of deep learning in medical imaging depends, in part, on the reliability of the predictions. Specifically, the system should be accurate for classes seen during training while providing calibrated estimates of uncertainty for abnormalities and unseen classes. To efficiently estimate predictive uncertainty, we propose the use of multi-head CNNs (M-heads). We compare its performance to related and more prevalent approaches, such as deep ensembles, on the task of out-of-distribution (OOD) detection. To this end, we evaluate models trained to discriminate normal lymph node tissue from breast cancer metastases, on lymph nodes containing lymphoma. We show the ability to discriminate between in-distribution lymph node tissue and lymphoma by evaluating the AUROC based on the uncertainty signal. Here, the best performing multi-head CNN (91.7) outperforms both Monte Carlo dropout (88.3) and deep ensembles (86.8). Furthermore, we show that the meta-loss function of M-heads improves OOD detection in terms of AUROC.} }
Endnote
%0 Conference Paper %T Efficient Out-of-Distribution Detection in Digital Pathology Using Multi-Head Convolutional Neural Networks %A Jasper Linmans %A Jeroen van der Laak %A Geert Litjens %B Proceedings of the Third Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2020 %E Tal Arbel %E Ismail Ben Ayed %E Marleen de Bruijne %E Maxime Descoteaux %E Herve Lombaert %E Christopher Pal %F pmlr-v121-linmans20a %I PMLR %P 465--478 %U https://proceedings.mlr.press/v121/linmans20a.html %V 121 %X Successful clinical implementation of deep learning in medical imaging depends, in part, on the reliability of the predictions. Specifically, the system should be accurate for classes seen during training while providing calibrated estimates of uncertainty for abnormalities and unseen classes. To efficiently estimate predictive uncertainty, we propose the use of multi-head CNNs (M-heads). We compare its performance to related and more prevalent approaches, such as deep ensembles, on the task of out-of-distribution (OOD) detection. To this end, we evaluate models trained to discriminate normal lymph node tissue from breast cancer metastases, on lymph nodes containing lymphoma. We show the ability to discriminate between in-distribution lymph node tissue and lymphoma by evaluating the AUROC based on the uncertainty signal. Here, the best performing multi-head CNN (91.7) outperforms both Monte Carlo dropout (88.3) and deep ensembles (86.8). Furthermore, we show that the meta-loss function of M-heads improves OOD detection in terms of AUROC.
APA
Linmans, J., van der Laak, J. & Litjens, G.. (2020). Efficient Out-of-Distribution Detection in Digital Pathology Using Multi-Head Convolutional Neural Networks. Proceedings of the Third Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 121:465-478 Available from https://proceedings.mlr.press/v121/linmans20a.html.

Related Material