Data Entry: Please note that the research database will be replaced by UNIverse by the end of October 2023. Please enter your data into the system https://universe-intern.unibas.ch. Thanks
Model-Driven 3-D Regularisation for Robust Segmentation of the Refractive Corneal Surfaces in Spiral {OCT} Scans
Book title (Conference Proceedings)
Fetal, Infant and Ophthalmic Medical Image Analysis
Place of Conference
Quebec
Publisher
Springer
Place of Publication
Cham
Pages
109-117
ISSN/ISBN
978-3-319-67560-2 ; 978-3-319-67561-9
Abstract
Measuring the cornea’s anterior and posterior refractive surface is essential for corneal topography, used for diagnostics and the planning of surgeries. Corneal topography by Optical Coherence Tomography (OCT) relies on proper segmentation. Common segmentation methods are limited to specific, B-scan-based scan patterns and fail when applied to data acquired by recently proposed spiral scan trajectories. We propose a novel method for the segmentation of the anterior and posterior refractive surface in scans acquired by 2-D scan trajectories – including but not limited to spirals. Key feature is a model-driven, three-dimensional regularisation of the region of interest, slope and curvature. The regularisation is integrated into a graph-based segmentation with feature-directed smoothing and incremental segmentation. We parameterise the segmentation based on test surface measurements and evaluate its performance by means of 18 in vivo measurements acquired by spiral and radial scanning. The comparison with expert segmentations shows successful segmentation of the refractive corneal surfaces.
Series title
Lecture Notes in Computer Science book series (LNCS)
