Visualizing the Palisades of Vogt: Limbal Registration by Surface Segmentation
Abstract
In this paper, a novel fully-automated algorithm is presented that enables the visualization of the Palisades of Vogt (POV) imaged using Optical Coherence Tomography (OCT). The algorithm segments the anterior limbal surface in each B-scan of a volume, uses it to guide the registration of individual B-scans to a reference B-scan, and creates an aligned 3D volume of the imaged limbal region. Accurate registration enables the extraction of cross-sections from below the ocular surface, thereby revealing intricate POV structural detail. Validation was performed across all imaged datasets. To the best of our knowledge, the proposed approach is the first fully automated approach to register the limbal surface of the eye for the express purpose of visualization of the POV.
BibTeX
@conference{Mathai and Galeotti and Lathrop-2018-109389,author = {Tejas Sudharshan Mathai and John Galeotti and Kira Lathrop},
title = {Visualizing the Palisades of Vogt: Limbal Registration by Surface Segmentation},
booktitle = {Proceedings of IEEE International Symposium on Biomedical Imaging (ISBI '18)},
year = {2018},
month = {April},
pages = {1327 - 1331},
publisher = {IEEE},
keywords = {Optical Coherence Tomography, Image segmentation , Motion Segmentation , 3D Reconstruction, Image Registration ,},
}