PuSH - Publikationsserver des Helmholtz Zentrums München

Wagner, S.J.* ; Khalili, N.* ; Sharma, R. ; Boxberg, M.* ; Marr, C. ; de Back, W.* ; Peng, T.

Structure-preserving multi-domain stain color augmentation using style-transfer with disentangled representations.

In: (24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021, 27 September-01 October 2021, Virtual, Online). Berlin [u.a.]: Springer, 2021. 257-266 (Lect. Notes Comput. Sc. ; 12908 LNCS)
DOI
In digital pathology, different staining procedures and scanners cause substantial color variations in whole-slide images (WSIs), especially across different laboratories. These color shifts result in a poor generalization of deep learning-based methods from the training domain to external pathology data. To increase test performance, stain normalization techniques are used to reduce the variance between training and test domain. Alternatively, color augmentation can be applied during training leading to a more robust model without the extra step of color normalization at test time. We propose a novel color augmentation technique, HistAuGAN, that can simulate a wide variety of realistic histology stain colors, thus making neural networks stain-invariant when applied during training. Based on a generative adversarial network (GAN) for image-to-image translation, our model disentangles the content of the image, i.e., the morphological tissue structure, from the stain color attributes. It can be trained on multiple domains and, therefore, learns to cover different stain colors as well as other domain-specific variations introduced in the slide preparation and imaging process. We demonstrate that HistAuGAN outperforms conventional color augmentation techniques on a classification task on the publicly available dataset Camelyon17 and show that it is able to mitigate present batch effects (Code and model weights are available at https://github.com/sophiajw/HistAuGAN.).
Altmetric
Weitere Metriken?
Zusatzinfos bearbeiten [➜Einloggen]
Publikationstyp Artikel: Konferenzbeitrag
Schlagwörter Color Augmentation ; Disentangled Representations ; Style-transfer
ISSN (print) / ISBN 0302-9743
e-ISSN 1611-3349
Konferenztitel 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
Konferzenzdatum 27 September-01 October 2021
Konferenzort Virtual, Online
Quellenangaben Band: 12908 LNCS, Heft: , Seiten: 257-266 Artikelnummer: , Supplement: ,
Verlag Springer
Verlagsort Berlin [u.a.]