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

• DOI: 10.48550/arxiv.2404.05409
• Publication Date: 2024-04-08
• Authors: Marc S. Seibel; Hristina Uzunova; Timo Kepp; Heinz Handels
• Keywords: FOS: Computer and information sciences; Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Computer Science - Computer Vision and Pattern Recognition; FOS: Electrical engineering, electronic engineering, information engineering; Electrical Engineering and Systems Science - Image and Video Processing

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) set Spectralis-OCT and Home-OCT images. I2I challenging because the are unpaired, bijective mapping does not exist due to information discrepancy between both domains. This has been addressed by Contrastive Learning for Unpaired Translation (CUT) approach, but it reduces semantic consistency. To restore consistency, we support style decoder additional segmentation decoder. Our approach increases similarity style-translated target distribution. Importantly, improve biomarkers in unsupervised domain adaptation scenario. provides potential monitoring diseases, e.g., age related macular disease, OCT devices.