Please use this identifier to cite or link to this item:
Title: Experimental study for 3D statistical property based intracranial brain tumor classification
Authors: Christe, S Allin
Kumari, B Vasantha
Kandaswamy, A
Keywords: Benign;Brain tumor;CAD;3D;Malignant;Statistical texture measures
Issue Date: Jan-2012
Publisher: NISCAIR-CSIR, India
Abstract: This study determines how 3D gray level co-occurrence statistical texture measures magnetic resonance (MR) images toprovide a better discrimination and classification of non-tumor/tumor tissues when compared to 2D texture measures using onlythe statistical properties and without using any conventional classifiers. Experiments, conducted with 50 sets of different nontumor/tumor volumes, demonstrate superiority of 3D texture characterization for tumor discrimination. Validation of tumordiscrimination was done, and performance measures achieved an overall accuracy (80%) with sensitivity of 85.71% and 83.33%,and specificity of 80.95% and 80.77% for non-tumor vs tumor and for benign vs malignant tumors, respectively. This study is apart of developing a computer aided diagnosis (CAD) system for classification of brain tumors in vivo before going for pathologicaltests like biopsy.
Page(s): 36-44
ISSN: 0975-1084 (Online); 0022-4456 (Print)
Appears in Collections:JSIR Vol.71(01) [January 2012]

Files in This Item:
File Description SizeFormat 
JSIR 71(1) 36-44.pdf273 kBAdobe PDFView/Open

Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.