Peer Reviewed

1

DOI

10.17077/omia.1016

Conference Location

Boston, MA, USA

Publication Date

September 2014

Abstract

Classify in vivo confocal microscopy corneal images by tortuosity is complicated by the presence of variable numbers of fibres of different tortuosity level. Instead of designing a function combining manually selected features into a single coefficient, as done in the literature, we propose a supervised approach which selects automatically the most relevant combination of shape features from a pre-defined dictionary. To our best knowledge, we are the first to consider features at different spatial scales and show experimentally their relevance in tortuosity modelling. Our results, obtained with a set of 100 images and 20 fold cross-validation, suggest that multinomial logistic ordinal regression, trained on consensus ground truth from 3 experts, yields an accuracy indistinguishable, overall, from that of experts when compared against each other.

Rights

Copyright © 2014, Roberto Annunziata, Ahmad Kheirkhah, Shruti Aggarwal, Bernardo M. Cavalcanti, Pedram Hamrah, and Emanuele Trucco.

Included in

Ophthalmology Commons

Share

COinS
 
Sep 14th, 12:00 AM Sep 14th, 12:00 AM

Tortuosity classification of corneal nerves images using a multiple-scale-multiple-window approach

Boston, MA, USA

Classify in vivo confocal microscopy corneal images by tortuosity is complicated by the presence of variable numbers of fibres of different tortuosity level. Instead of designing a function combining manually selected features into a single coefficient, as done in the literature, we propose a supervised approach which selects automatically the most relevant combination of shape features from a pre-defined dictionary. To our best knowledge, we are the first to consider features at different spatial scales and show experimentally their relevance in tortuosity modelling. Our results, obtained with a set of 100 images and 20 fold cross-validation, suggest that multinomial logistic ordinal regression, trained on consensus ground truth from 3 experts, yields an accuracy indistinguishable, overall, from that of experts when compared against each other.