Paper Abstract and Keywords |
Presentation |
2006-01-27 17:45
Improvement of simultaneous segmentation of multi-organ based on estimation of feature distribution parameters Rena Ohno, Hironori Sakurai (TUAT), Daniel Smutek (Charles Univ.), Akinobu Shimizu, Hidefumi Kobatake (TUAT), Shigeru Nawano (National Cancer Center Hospital East) |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we present a simultaneous segmentation method for multi-organ in three dimensional abdominal CT images based on digital atlas of human anatomy and EM algorithm. In this method, feature distribution parameters (mean, covariance) are estimated by EM algorithm. We extend the mixture distribution model for EM estimation so that weights (mixture ratios) for distributions can be defined at each voxel in EM algorithm. To validate the proposed algorithm, we apply it to actual 10 3D abdominal CT images and evaluate the accuracy of parameter estimation and simultaneous segmentation. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
3D CT image / Multi-organ simultaneous segmentation / EM Algorithm / Probability atlas / Maximum a posterior / / / |
Reference Info. |
IEICE Tech. Rep., vol. 105, no. 579, MI2005-106, pp. 159-162, Jan. 2006. |
Paper # |
MI2005-106 |
Date of Issue |
2006-01-20 (MI) |
ISSN |
Print edition: ISSN 0913-5685 |
Download PDF |
|