Paper Abstract and Keywords |
Presentation |
2018-07-24 10:15
Classification of Corneal Shape Images by Machine Learning Yuji Ayatsuka (CRESCO), Kazutaka Kamiya (Kitasato Univ.), Yudai Kato, Yusuke Kudo (CRESCO), Fusako Fujimura, Nobuyuki Shoji (Kitasato Univ.), Yosai Mori, Kazunori Miyata (Miyata Eye Hospital) MI2018-22 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Keratoconus is one of an ophthalmic disease on a cornea, which brings thinning and deformation in the center of cornea and causes high myopia or astigmatism. Anterior segment optical coherence tomography (Anterior Segment OCT) is used to diagnose keratoconus including its stage. We applied machine learning to analyze images by anterior segment OCT. A resulted trained model, based on convolutional neural network, classified eyes of normal and grade 1 to 4 in Amsler-Krumeich classification with 85.3% accuracy. It also classified normal eyes and keratoconus eyes with 99.6% accuracy. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
keratoconus / Amsler-Krumeich classification / machine learning / anterior segment optical coherence tomography / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 118, no. 150, MI2018-22, pp. 5-8, July 2018. |
Paper # |
MI2018-22 |
Date of Issue |
2018-07-17 (MI) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
MI2018-22 |
|