Home // ICCGI 2013, The Eighth International Multi-Conference on Computing in the Global Information Technology // View article

Actual Brain Midline Detection using Level Set Segmentation and Window Selection

Authors:
Xuguang Qi
Ashwin Belle
Sharad Shandilya
Rosalyn Hargraves
Charles Cockrell
Yang Tang
Kevin Ward
Kayvan Najarian

Keywords: actual midline; ventricle; CT slice; window selection; level set segmentation

Abstract:
Detection of actual brain midline is essential for accurate estimation of midline shift due to traumatic brain injury. An effective method to estimate the actual midline is to use the positions of the identified ventricles. In this work, a level set algorithm combined with Ventricle Window Selection Algorithm and Ventricle Identification Algorithm is proposed to detect the ventricular system in computed tomography (CT) images. The system automatically selects appropriate slices from numerous raw slices and confines the focus to the region of interest prior to segmentation. The variational level set method performs ventricle segmentation without any requirement of re-initialization or intensity-homogeneity of CT images. Combined with ventricle identification, the level set segmentation successfully extracts ventricle contours and estimates actual midline. Experimental results assessed on 391 CT slices of 40 patients support that the proposed system is accurate (90%) and useful for clinical practice.

Pages: 122 to 126

Copyright: Copyright (c) IARIA, 2013

Publication date: July 21, 2013

Published in: conference

ISSN: 2308-4529

ISBN: 978-1-61208-283-7

Location: Nice, France

Dates: from July 21, 2013 to July 26, 2013