Automated torso organ segmentation from 3D CT images using structured perceptron and dual decomposition

Yukitaka Nimura; Yuichiro Hayashi; Takayuki Kitasaka; Kensaku Mori

doi:10.1117/12.2081774

20 March 2015 Automated torso organ segmentation from 3D CT images using structured perceptron and dual decomposition

Yukitaka Nimura, Yuichiro Hayashi, Takayuki Kitasaka, Kensaku Mori

Proceedings Volume 9414, Medical Imaging 2015: Computer-Aided Diagnosis; 94143L (2015) https://doi.org/10.1117/12.2081774
Event: SPIE Medical Imaging, 2015, Orlando, Florida, United States

Abstract

This paper presents a method for torso organ segmentation from abdominal CT images using structured perceptron and dual decomposition. A lot of methods have been proposed to enable automated extraction of organ regions from volumetric medical images. However, it is necessary to adjust empirical parameters of them to obtain precise organ regions. This paper proposes an organ segmentation method using structured output learning. Our method utilizes a graphical model and binary features which represent the relationship between voxel intensities and organ labels. Also we optimize the weights of the graphical model by structured perceptron and estimate the best organ label for a given image by dynamic programming and dual decomposition. The experimental result revealed that the proposed method can extract organ regions automatically using structured output learning. The error of organ label estimation was 4.4%. The DICE coeﬃcients of left lung, right lung, heart, liver, spleen, pancreas, left kidney, right kidney, and gallbladder were 0.91, 0.95, 0.77, 0.81, 0.74, 0.08, 0.83, 0.84, and 0.03, respectively.

Citation Download Citation

Yukitaka Nimura, Yuichiro Hayashi, Takayuki Kitasaka, and Kensaku Mori "Automated torso organ segmentation from 3D CT images using structured perceptron and dual decomposition", Proc. SPIE 9414, Medical Imaging 2015: Computer-Aided Diagnosis, 94143L (20 March 2015); https://doi.org/10.1117/12.2081774

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

Includes PDF, HTML & Video, when available