Flash News
Welcome to IARC

Publisher Login

Latest News
Welcome to IARC- JCR Report

Submit your Journal to get IARC-JCRR Indexing and Impact Factor
 

Impact Factor calculated by IARC on the basis of Journal Citation Reference (JCR) Report.

 

Contact: iarcdpi@gmail.com

 

OBL-GA based FCM with level sets for automatic GBM tumor segmentation in MR Images
B. Srinivasa Rao and E. Sreenivasa Reddy

Published in: International Journal of Computer Sciences and Engineering
Volume- 5, Issue-1, pp.85-20, Jan 2017
DPI :-> 16.10022.IJCSE.2017.V5I1.8520.1078



Abstract
This paper presents an automatic method for the segmentation of Glioblastoma multiforme(GBM) tumors from MRI images. The global search ability of Genetic Algorithm (GA) to optimize the Fuzzy C-means (FCM) clustering algorithm to obtain better clustering center. But the prematurity problem of GA itself has bad effects on the whole clustering. Therefore, in order to optimize the traditional GA-FCM algorithm’s clustering effect, in this work, we introduce the Opposition-based learning mechanism into GA, to construct an OBL-Genetic Algorithm (OBL-GA). The improved algorithm forms the next generation of evolutionary population by selecting the superior individuals in the collection of the sub generation and reverse sub generation, to increase the population diversity, and final to overcome the prematurity problem of GA. Then applying the improved algorithm to FCM, which gives better results and then resultant image, is applied with level sets, to exact delineation of GBM tumor. The validation is performed on a labeled BRATS data set. Our segmentation results are highly accurate, and compare favorably to the state of the art.

Key-Words / Index Term
Fuzzy-c means,Glioblastoma multiforme,Segmentation,Genetic Algorithm,Opposition based learning,MRI

How to cite this article
B. Srinivasa Rao and E. Sreenivasa Reddy , “OBL-GA based FCM with level sets for automatic GBM tumor segmentation in MR Images”, International Journal of Computer Sciences and Engineering, 5, Issue-1, pp.85-20, Jan 2017. DPI:16.10022.IJCSE.V5.I1.1078