A Vesselsegmentation-based CycleGAN for Unpaired Multi-modal Retinal Image Synthesis

Sindel A, Maier A, Christlein V (2023)


Publication Language: English

Publication Type: Conference contribution, Conference Contribution

Publication year: 2023

Publisher: Springer Vieweg

City/Town: Wiesbaden

Pages Range: 32-37

Conference Proceedings Title: Bildverarbeitung für die Medizin 2023

Event location: Braunschweig

URI: https://arxiv.org/pdf/2306.02901.pdf

DOI: 10.1007/978-3-658-41657-7_11

Abstract

Unpaired image-to-image translation of retinal images can efficiently increase the training dataset for deep-learning-based multi-modal retinal registration methods. Our method integrates a vessel segmentation network into the image-to-image translation task by extending the CycleGAN framework. The segmentation network is inserted prior to aUNet vision transformer generator network and serves as a shared representation between both domains. We reformulate the original identity loss to learn the direct mapping between the vessel segmentation and the real image. Additionally, we add a segmentation loss termto ensure shared vessel locations between fake and real images. In the experiments, our method shows a visually realistic look and preserves the vessel structures, which is a prerequisite for generating multi-modal training data for image registration.

Authors with CRIS profile

How to cite

APA:

Sindel, A., Maier, A., & Christlein, V. (2023). A Vesselsegmentation-based CycleGAN for Unpaired Multi-modal Retinal Image Synthesis. In Bildverarbeitung für die Medizin 2023 (pp. 32-37). Braunschweig: Wiesbaden: Springer Vieweg.

MLA:

Sindel, Aline, Andreas Maier, and Vincent Christlein. "A Vesselsegmentation-based CycleGAN for Unpaired Multi-modal Retinal Image Synthesis." Proceedings of the Bildverarbeitung für die Medizin 2023, Braunschweig Wiesbaden: Springer Vieweg, 2023. 32-37.

BibTeX: Download