Automatic anatomy partitioning of the torso region on CT images by using multiple organ localizations with a group-wise calibration technique

Xiangrong Zhou; Syoichi Morita; Xinxin Zhou; Huayue Chen; Takeshi Hara; Ryujiro Yokoyama; Masayuki Kanematsu; Hiroaki Hoshi; Hiroshi Fujita

doi:10.1117/12.2081786

20 March 2015 Automatic anatomy partitioning of the torso region on CT images by using multiple organ localizations with a group-wise calibration technique

Xiangrong Zhou, Syoichi Morita, Xinxin Zhou, Huayue Chen, Takeshi Hara, Ryujiro Yokoyama, Masayuki Kanematsu, Hiroaki Hoshi, Hiroshi Fujita

Author Affiliations +

Proceedings Volume 9414, Medical Imaging 2015: Computer-Aided Diagnosis; 94143K (2015) https://doi.org/10.1117/12.2081786
Event: SPIE Medical Imaging, 2015, Orlando, Florida, United States

Abstract

This paper describes an automatic approach for anatomy partitioning on three-dimensional (3D) computedtomography (CT) images that divide the human torso into several volume-of-interesting (VOI) images based on anatomical definition. The proposed approach combines several individual detections of organ-location with a groupwise organ-location calibration and correction to achieve an automatic and robust multiple-organ localization task. The essence of the proposed method is to jointly detect the 3D minimum bounding box for each type of organ shown on CT images based on intra-organ-image-textures and inter-organ-spatial-relationship in the anatomy. Machine-learning-based template matching and generalized Hough transform-based point-distribution estimation are used in the detection and calibration processes. We apply this approach to the automatic partitioning of a torso region on CT images, which are divided into 35 VOIs presenting major organ regions and tissues required by routine diagnosis in clinical medicine. A database containing 4,300 patient cases of high-resolution 3D torso CT images is used for training and performance evaluations. We confirmed that the proposed method was successful in target organ localization on more than 95% of CT cases. Only two organs (gallbladder and pancreas) showed a lower success rate: 71 and 78% respectively. In addition, we applied this approach to another database that included 287 patient cases of whole-body CT images scanned for positron emission tomography (PET) studies and used for additional performance evaluation. The experimental results showed that no significant difference between the anatomy partitioning results from those two databases except regarding the spleen. All experimental results showed that the proposed approach was efficient and useful in accomplishing localization tasks for major organs and tissues on CT images scanned using different protocols.

Citation Download Citation

Xiangrong Zhou, Syoichi Morita, Xinxin Zhou, Huayue Chen, Takeshi Hara, Ryujiro Yokoyama, Masayuki Kanematsu, Hiroaki Hoshi, and Hiroshi Fujita "Automatic anatomy partitioning of the torso region on CT images by using multiple organ localizations with a group-wise calibration technique", Proc. SPIE 9414, Medical Imaging 2015: Computer-Aided Diagnosis, 94143K (20 March 2015); https://doi.org/10.1117/12.2081786

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE