Fully-automatic multimodal co-registration of retinal fundus images

Abstract

Optical coherence tomography (OCT), color fundus photography (CFP), and fluorescein angiography (FA) are imaging modalities widely used in ophthalmology, and their coregistration is of major interest. The integrated overview ofmultiple images of the human retina may open new perspectives for the understanding of changes in the human eye in health and disease, thus improving our knowledge on retinal diagnostics. We propose a fast fully-automatic algorithm to co-register retinal fundus images from multiple modalities. Our results show that the herewith proposed algorithm is able to work with distinct levels of vascular network detail, contrast, and resolution across OCT, CFP, and FA imaging. Furthermore, the achieved processing time (1.7 seconds) makes it an asset for the daily clinical practice.