| IEICE Technical Committee Submission System Conference Paper's Information |
Online Proceedings [Sign in] Tech. Rep. Archives |
| Go Top Page | Go Previous | [Japanese] / [English] |
| Paper Abstract and Keywords | ||
| Presentation | 2022-01-27 16:39 [Short Paper] Multiple Organ Detection from CT Images Based on Deep Learning -- Fusion of 2D-CNN and Transformer -- Daiki Kanoh, Xiangrong Zhou, Takeshi Hara, Hiroshi Fujita (Gifu Univ.)  MI2021-89 |
|
| Abstract | (in Japanese) | (See Japanese page) |
| (in English) | The automatic recognition of multiple organs in 3D CT images and detecting organ positions are required for computer-aided diagnosis systems to support doctors' diagnosis. This study aims to automatically recognize and localize multiple organs on 3D CT images by using deep learning approach. In our previous study, we proposed a method to integrate organ locations by detecting 2D cross-sectional locations and vote those 2D locations in a 3D space. Because this method did not directly use 3D image features, it should be a room for further improvement by using a 3D feature map. In this study, we propose an end-to-end 3D localization method using a transformer with a 3D feature map generated from 2D-CNN on CT images that covered a wide range of the human body. The proposed method is a modification of the Spine-Transformers proposed by a previous work. In our previous method, we used an improved SSD. We applied each method to 17 types of organs and evaluated the localization results. For training and testing, we used 240 cases from a shared dataset from a research project "Computational Anatomy", which contains a mixture of contrast and non-contrast CT images. The preliminary results showed that the proposed method could not performed a high accuracy for multiple organ localizations over a wide range of CT images. From the performance comparison between the proposed method based on 3D features and our previous method based on 2D features, we confirmed that our previous approach, "majority voting in 3D based on organ localization on 2D images," is still a realistic and effective method for multiple organ localizations on 3D CT images. | |
| Keyword | (in Japanese) | (See Japanese page) |
| (in English) | 3D CT Image / multi organ detection / Transformer / feature aggregation / / / / | |
| Reference Info. | IEICE Tech. Rep., vol. 121, no. 347, MI2021-89, pp. 190-193, Jan. 2022. | |
| Paper # | MI2021-89 | |
| Date of Issue | 2022-01-18 (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 | MI2021-89 |
|
| Conference Information | |
| Committee | MI |
| Conference Date | 2022-01-25 - 2022-01-27 |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Paper Information | |
| Registration To | MI |
| Conference Code | 2022-01-MI |
| Language | Japanese |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Multiple Organ Detection from CT Images Based on Deep Learning |
| Sub Title (in English) | Fusion of 2D-CNN and Transformer |
| Keyword(1) | 3D CT Image |
| Keyword(2) | multi organ detection |
| Keyword(3) | Transformer |
| Keyword(4) | feature aggregation |
| Keyword(5) | |
| Keyword(6) | |
| Keyword(7) | |
| Keyword(8) | |
| 1st Author's Name | Daiki Kanoh |
| 1st Author's Affiliation | Gifu University (Gifu Univ.) |
| 2nd Author's Name | Xiangrong Zhou |
| 2nd Author's Affiliation | Gifu University (Gifu Univ.) |
| 3rd Author's Name | Takeshi Hara |
| 3rd Author's Affiliation | Gifu University (Gifu Univ.) |
| 4th Author's Name | Hiroshi Fujita |
| 4th Author's Affiliation | Gifu University (Gifu Univ.) |
| 5th Author's Name | |
| 5th Author's Affiliation | () |
| 6th Author's Name | |
| 6th Author's Affiliation | () |
| 7th Author's Name | |
| 7th Author's Affiliation | () |
| 8th Author's Name | |
| 8th Author's Affiliation | () |
| 9th Author's Name | |
| 9th Author's Affiliation | () |
| 10th Author's Name | |
| 10th Author's Affiliation | () |
| 11th Author's Name | |
| 11th Author's Affiliation | () |
| 12th Author's Name | |
| 12th Author's Affiliation | () |
| 13th Author's Name | |
| 13th Author's Affiliation | () |
| 14th Author's Name | |
| 14th Author's Affiliation | () |
| 15th Author's Name | |
| 15th Author's Affiliation | () |
| 16th Author's Name | |
| 16th Author's Affiliation | () |
| 17th Author's Name | |
| 17th Author's Affiliation | () |
| 18th Author's Name | |
| 18th Author's Affiliation | () |
| Speaker | Author-1 |
| Date Time | 2022-01-27 16:39:00 |
| Presentation Time | 13 minutes |
| Registration for | MI |
| Paper # | MI2021-89 |
| Volume (vol) | vol.121 |
| Number (no) | no.347 |
| Page | pp.190-193 |
| #Pages | 4 |
| Date of Issue | 2022-01-18 (MI) |
[Return to Top Page]