Paper Abstract and Keywords |
Presentation |
2009-01-21 09:50
[Poster Presentation]
Construction of Hierarchical Statistical Shape and Motion Model of Hip Joint Futoshi Yokota (Kobe Univ.), Toshiyuki Okada, Masahiko Nakamoto, Masaki Takao, Nobuhiko Sugano, Hideki Yoshikawa (Osaka Univ.), Yukio Tada (Kobe Univ.), Yoshinobu Sato (Osaka Univ.) MI2008-180 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
Segmentation and separation of the femur and pelvis from 3D CT images are prerequisite of patient specific surgical planning and simulation. Separation of the femoral head and acetablum is one of main difficulties for diseased hip joint because the joint space becomes extremely narrow. Our current method for the separation involves manual specification using independent femur and pelvis statistical shape models (SSMs). In this paper, we construct and evaluate statistical shape and motion model (SSMM) of the hip joint, in which both the femur and pelvis are integrated, to automate the separation. Possible joint poses are simulationally generated and embedded into SSMM to deal with unknown arrangement of the femur and pelvis in the CT images. Further, the SSMM is hierarchically constructed to represent wider variety of femur and pelvis shapes for different poses. In the experiments, SSMM of the hip joint was constructed using 22 cases. We evaluated prediction performance of unknown femoral head and acetabulum shapes from known other parts of the femur and pelvis using the proposed SSMM, and confirmed accuracy improvement of around 25% compared with conventional independent SSMs. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
multi-organ statistical shape model / hierarchical statistical shape model / hip joint / joint motion / segmentation / / / |
Reference Info. |
IEICE Tech. Rep., vol. 108, no. 385, MI2008-180, pp. 545-550, Jan. 2009. |
Paper # |
MI2008-180 |
Date of Issue |
2009-01-12 (MI) |
ISSN |
Print edition: ISSN 0913-5685 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 |
MI2008-180 |
|