Paper Abstract and Keywords |
Presentation |
2021-03-15 13:45
Surgical planning model generation by extracting important feature sets in mandibular reconstruction Kazuki Nagai, Megumi Nakao (Kyoto Univ.), Nobuhiro Ueda (Nara Medical Univ.), Yuichiro Imai (Rakuwakai Otowa Hospital), Toshihide Hatanaka, Tadaaki Kirita (Nara Medical Univ.), Tetsuya Matsuda (Kyoto Univ.) MI2020-54 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Because implicit medical knowledge and experience are used to perform medical treatment, such decisions must be clarified when systematizing surgical procedures. We proposed the enumaration of Lasso solutions corresponding to multiple classes in mandibular reconstruction, but the calculation amount required to search feature sets was large. In this study, we propose the enumeration of Lasso solutions algorithm for multi-class classification that preferentioally selects frequently features. Experiments showed that the 5-dimensional feature set which can correctly estimate more than 90% of surgeons' plans with 76% calculation time compared to the previous methods. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Enumeration of Lasso solutions / Interpretability / Feature extraction / Mandibular reconstruction / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 120, no. 431, MI2020-54, pp. 29-34, March 2021. |
Paper # |
MI2020-54 |
Date of Issue |
2021-03-08 (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 |
MI2020-54 |
|