Anatomical Conditioning for Contrastive Unpaired Image-to-Image Translation of Optical Coherence Tomography Images

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Seibel, Marc S., Uzunova, Hristina, Kepp, Timo, Handels, Heinz
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929305650462720
author Seibel, Marc S.
Uzunova, Hristina
Kepp, Timo
Handels, Heinz
author_facet Seibel, Marc S.
Uzunova, Hristina
Kepp, Timo
Handels, Heinz
contents For a unified analysis of medical images from different modalities, data harmonization using image-to-image (I2I) translation is desired. We study this problem employing an optical coherence tomography (OCT) data set of Spectralis-OCT and Home-OCT images. I2I translation is challenging because the images are unpaired, and a bijective mapping does not exist due to the information discrepancy between both domains. This problem has been addressed by the Contrastive Learning for Unpaired I2I Translation (CUT) approach, but it reduces semantic consistency. To restore the semantic consistency, we support the style decoder using an additional segmentation decoder. Our approach increases the similarity between the style-translated images and the target distribution. Importantly, we improve the segmentation of biomarkers in Home-OCT images in an unsupervised domain adaptation scenario. Our data harmonization approach provides potential for the monitoring of diseases, e.g., age related macular disease, using different OCT devices.
format Preprint
id arxiv_https___arxiv_org_abs_2404_05409
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Anatomical Conditioning for Contrastive Unpaired Image-to-Image Translation of Optical Coherence Tomography Images
Seibel, Marc S.
Uzunova, Hristina
Kepp, Timo
Handels, Heinz
Image and Video Processing
Computer Vision and Pattern Recognition
For a unified analysis of medical images from different modalities, data harmonization using image-to-image (I2I) translation is desired. We study this problem employing an optical coherence tomography (OCT) data set of Spectralis-OCT and Home-OCT images. I2I translation is challenging because the images are unpaired, and a bijective mapping does not exist due to the information discrepancy between both domains. This problem has been addressed by the Contrastive Learning for Unpaired I2I Translation (CUT) approach, but it reduces semantic consistency. To restore the semantic consistency, we support the style decoder using an additional segmentation decoder. Our approach increases the similarity between the style-translated images and the target distribution. Importantly, we improve the segmentation of biomarkers in Home-OCT images in an unsupervised domain adaptation scenario. Our data harmonization approach provides potential for the monitoring of diseases, e.g., age related macular disease, using different OCT devices.
title Anatomical Conditioning for Contrastive Unpaired Image-to-Image Translation of Optical Coherence Tomography Images
topic Image and Video Processing
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2404.05409