Extending Unsupervised Neural Image Compression With Supervised Multitask Learning

David Tellez, Diederik Höppener, Cornelis Verhoef, Dirk Grünhagen, Pieter Nierop, Michal Drozdzal, Jeroen Laak, Francesco Ciompi
Proceedings of the Third Conference on Medical Imaging with Deep Learning, PMLR 121:770-783, 2020.

Abstract

We focus on the problem of training convolutional neural networks on gigapixel histopathology images to predict image-level targets. For this purpose, we extend Neural Image Compression (NIC), an image compression framework that reduces the dimensionality of these images using an encoder network trained unsupervisedly. We propose to train this encoder using supervised multitask learning (MTL) instead. We applied the proposed MTL NIC to two histopathology datasets and three tasks. First, we obtained state-of-the-art results in the Tumor Proliferation Assessment Challenge of 2016 (TUPAC16). Second, we successfully classified histopathological growth patterns in images with colorectal liver metastasis (CLM). Third, we predicted patient risk of death by learning directly from overall survival in the same CLM data. Our experimental results suggest that the representations learned by the MTL objective are: (1) highly specific, due to the supervised training signal, and (2) transferable, since the same features perform well across different tasks. Additionally, we trained multiple encoders with different training objectives, e.g. unsupervised and variants of MTL, and observed a positive correlation between the number of tasks in MTL and the system performance on the TUPAC16 dataset.

Cite this Paper

BibTeX
@InProceedings{pmlr-v121-tellez20a, title = {Extending Unsupervised Neural Image Compression With Supervised Multitask Learning}, author = {Tellez, David and H\"oppener, Diederik and Verhoef, Cornelis and Gr\"unhagen, Dirk and Nierop, Pieter and Drozdzal, Michal and van der Laak, Jeroen and Ciompi, Francesco}, booktitle = {Proceedings of the Third Conference on Medical Imaging with Deep Learning}, pages = {770--783}, 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/tellez20a/tellez20a.pdf}, url = {https://proceedings.mlr.press/v121/tellez20a.html}, abstract = {We focus on the problem of training convolutional neural networks on gigapixel histopathology images to predict image-level targets. For this purpose, we extend Neural Image Compression (NIC), an image compression framework that reduces the dimensionality of these images using an encoder network trained unsupervisedly. We propose to train this encoder using supervised multitask learning (MTL) instead. We applied the proposed MTL NIC to two histopathology datasets and three tasks. First, we obtained state-of-the-art results in the Tumor Proliferation Assessment Challenge of 2016 (TUPAC16). Second, we successfully classified histopathological growth patterns in images with colorectal liver metastasis (CLM). Third, we predicted patient risk of death by learning directly from overall survival in the same CLM data. Our experimental results suggest that the representations learned by the MTL objective are: (1) highly specific, due to the supervised training signal, and (2) transferable, since the same features perform well across different tasks. Additionally, we trained multiple encoders with different training objectives, e.g. unsupervised and variants of MTL, and observed a positive correlation between the number of tasks in MTL and the system performance on the TUPAC16 dataset.} }
Endnote
%0 Conference Paper %T Extending Unsupervised Neural Image Compression With Supervised Multitask Learning %A David Tellez %A Diederik Höppener %A Cornelis Verhoef %A Dirk Grünhagen %A Pieter Nierop %A Michal Drozdzal %A Jeroen Laak %A Francesco Ciompi %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-tellez20a %I PMLR %P 770--783 %U https://proceedings.mlr.press/v121/tellez20a.html %V 121 %X We focus on the problem of training convolutional neural networks on gigapixel histopathology images to predict image-level targets. For this purpose, we extend Neural Image Compression (NIC), an image compression framework that reduces the dimensionality of these images using an encoder network trained unsupervisedly. We propose to train this encoder using supervised multitask learning (MTL) instead. We applied the proposed MTL NIC to two histopathology datasets and three tasks. First, we obtained state-of-the-art results in the Tumor Proliferation Assessment Challenge of 2016 (TUPAC16). Second, we successfully classified histopathological growth patterns in images with colorectal liver metastasis (CLM). Third, we predicted patient risk of death by learning directly from overall survival in the same CLM data. Our experimental results suggest that the representations learned by the MTL objective are: (1) highly specific, due to the supervised training signal, and (2) transferable, since the same features perform well across different tasks. Additionally, we trained multiple encoders with different training objectives, e.g. unsupervised and variants of MTL, and observed a positive correlation between the number of tasks in MTL and the system performance on the TUPAC16 dataset.
APA
Tellez, D., Höppener, D., Verhoef, C., Grünhagen, D., Nierop, P., Drozdzal, M., Laak, J. & Ciompi, F.. (2020). Extending Unsupervised Neural Image Compression With Supervised Multitask Learning. Proceedings of the Third Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 121:770-783 Available from https://proceedings.mlr.press/v121/tellez20a.html.

Related Material