Paper Abstract and Keywords |
Presentation |
2020-01-29 13:20
[Short Paper]
Recognition of dentition using deep learning in dental panoramic X-ray images Takumi Morishita (Gifu Univ.), Chisako Muramatsu (Shiga Univ.), Xiangrong Zhou (Gifu Univ.), Ryo Takahashi, Tatsuro Hayashi (media), Wataru Nishiyama (Asahi Univ), Takeshi Hara (Gifu Univ.), Yoshiko Ariji, Eiichiro Ariji (Aichi Gakuin Univ.), Akitoshi Katsumata (Asahi Univ), Hiroshi Fujita (Gifu Univ.) MI2019-81 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Dental panoramic X-ray images are obtained at more than 98% of dental clinics in Japan. However, since all the teeth are depicted in the image, it takes a lot of time and effort for dentists to make an accurate image diagnosis of all teeth. The purpose of this study is to analyze dental panoramic X-ray images using deep learning, and to contribute to the diagnosis by dentists. As the initial stage, we detected teeth in images and classified their tooth types. We compared the accuracy of the Single Shot Multibox Detector (SSD) with that of the proposed method based on the SSD with a network branch. As a result, the proposed method exceeded the original SSD. The detection rate of the teeth was 96.5%, and the number of false detections was 0.32. The classification rate was 95.5% for 14 tooth types. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Dental panoramic X-ray images / Deep Learning / CAD / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 399, MI2019-81, pp. 73-74, Jan. 2020. |
Paper # |
MI2019-81 |
Date of Issue |
2020-01-22 (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 |
MI2019-81 |
|