Paper Abstract and Keywords |
Presentation |
2020-01-29 10:15
Analysis of disease classification and musculoskeletal anatomy using medical images and radiology reports in a large-scale medical image database Shuhei Honda, Yoshito Otake (NAIST), Masaki Takao (Osaka Univ.), Eiji Aramaki, Shuntaro Yada, Yuta Hiasa (NAIST), Kento Aida, Shinichi Sato (NII), Akihiro Nishie (Kyushu Univ.), Nobuhiko Sugano (Osaka Univ.), Yoshinobu Sato (NAIST) MI2019-69 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Recently, the environment for the analysis of large databases, such as the large-scale medical image database, have been constructed. Due to this, our research group has been developed an automatic analysis system to analyze musculoskeletal anatomical parameters by gender and age using convolutional neural network (CNN). However, patient-specific backgrounds such as disease are not known, so it is necessary to clarify patient-specific backgrounds to acquire new medical knowledge. In this study, in order to clarify the patient-specific background, a radiology report created by a radiologist when CT images were taken was analyzed by natural language processing to classify diseases. In addition, we analyze the relationship between musculoskeletal anatomical parameters and disease classification, and attempt to acquire new medical knowledge. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
CT Image / Radiology Report / NLP / CNN / BERT / Medical Image Database / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 399, MI2019-69, pp. 19-22, Jan. 2020. |
Paper # |
MI2019-69 |
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 |
MI2019-69 |
|