Domain Generalization for Retinal Vessel Segmentation with Vector Field Transformer

Dewei Hu, Hao Li, Han Liu, Ipek Oguz
Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, PMLR 172:552-564, 2022.

Abstract

Domain generalization has great impact on medical image analysis as data distribution inconsistencies are prevalent in most of the medical data modalities due to the image acquisition techniques. In this study, we investigate a novel pipeline that generalizes the retinal vessel segmentation across color fundus photography and OCT angiography images. We hypothesize that the scaled minor eigenvector of the Hessian matrix can sufficiently represent the vessel by vector flow. This vector field can be regarded as a common domain for different modalities as it is very similar even for data that follows vastly different intensity distributions. Next, we leverage the uncertainty in the latent space of the auto-encoder to synthesize enhanced vessel maps to augment the training data. Finally, we propose a transformer network to extract features from the vector field. We show the performance of our model in cross-modality experiments.

Cite this Paper

BibTeX
@InProceedings{pmlr-v172-hu22a, title = {Domain Generalization for Retinal Vessel Segmentation with Vector Field Transformer}, author = {Hu, Dewei and Li, Hao and Liu, Han and Oguz, Ipek}, booktitle = {Proceedings of The 5th International Conference on Medical Imaging with Deep Learning}, pages = {552--564}, year = {2022}, editor = {Konukoglu, Ender and Menze, Bjoern and Venkataraman, Archana and Baumgartner, Christian and Dou, Qi and Albarqouni, Shadi}, volume = {172}, series = {Proceedings of Machine Learning Research}, month = {06--08 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v172/hu22a/hu22a.pdf}, url = {https://proceedings.mlr.press/v172/hu22a.html}, abstract = {Domain generalization has great impact on medical image analysis as data distribution inconsistencies are prevalent in most of the medical data modalities due to the image acquisition techniques. In this study, we investigate a novel pipeline that generalizes the retinal vessel segmentation across color fundus photography and OCT angiography images. We hypothesize that the scaled minor eigenvector of the Hessian matrix can sufficiently represent the vessel by vector flow. This vector field can be regarded as a common domain for different modalities as it is very similar even for data that follows vastly different intensity distributions. Next, we leverage the uncertainty in the latent space of the auto-encoder to synthesize enhanced vessel maps to augment the training data. Finally, we propose a transformer network to extract features from the vector field. We show the performance of our model in cross-modality experiments.} }
Endnote
%0 Conference Paper %T Domain Generalization for Retinal Vessel Segmentation with Vector Field Transformer %A Dewei Hu %A Hao Li %A Han Liu %A Ipek Oguz %B Proceedings of The 5th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2022 %E Ender Konukoglu %E Bjoern Menze %E Archana Venkataraman %E Christian Baumgartner %E Qi Dou %E Shadi Albarqouni %F pmlr-v172-hu22a %I PMLR %P 552--564 %U https://proceedings.mlr.press/v172/hu22a.html %V 172 %X Domain generalization has great impact on medical image analysis as data distribution inconsistencies are prevalent in most of the medical data modalities due to the image acquisition techniques. In this study, we investigate a novel pipeline that generalizes the retinal vessel segmentation across color fundus photography and OCT angiography images. We hypothesize that the scaled minor eigenvector of the Hessian matrix can sufficiently represent the vessel by vector flow. This vector field can be regarded as a common domain for different modalities as it is very similar even for data that follows vastly different intensity distributions. Next, we leverage the uncertainty in the latent space of the auto-encoder to synthesize enhanced vessel maps to augment the training data. Finally, we propose a transformer network to extract features from the vector field. We show the performance of our model in cross-modality experiments.
APA
Hu, D., Li, H., Liu, H. & Oguz, I.. (2022). Domain Generalization for Retinal Vessel Segmentation with Vector Field Transformer. Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 172:552-564 Available from https://proceedings.mlr.press/v172/hu22a.html.

Related Material