Please use this identifier to cite or link to this item: http://nopr.niscair.res.in/handle/123456789/55420
Title: Hybrid Techniques for MRI Spine Images Classification
Authors: Raja, Geetha
Mohan, J
Keywords: CNN;KNN;LBP;PCA;SVM
Issue Date: Sep-2020
Publisher: NISCAIR-CSIR, India
Abstract: The number of persons suffering from spinal tumor has increased significantly from 2010 to 2016. Tumor is one of the major diseases of spinal cord. Thousands of researchers have concentrated on this disease to provide more efficient diagnosis with better understanding of the classification of spinal cord tumor. The proposed convolutional neural network (CNN) is tested with two hybrid recognized techniques of image detection which are K–nearest neighbor (KNN) with principal component analysis (PCA), local binary patterns (LBP) with support vector machine (SVM). Above three techniques overall accuracy is demonstrated, which show that LBP with SVM gives better result compared to KNN with PCA. The proposed CNN provides high accuracy classification and detection of spine diseases compared to other three techniques, which have obtained a best detection accuracy of 99.41 %. This process is fully implemented in MATLAB tool.
Page(s): 829-832
URI: http://nopr.niscair.res.in/handle/123456789/55420
ISSN: 0975-1084 (Online); 0022-4456 (Print)
Appears in Collections:JSIR Vol.79(09) [September 2020]

Files in This Item:
File Description SizeFormat 
JSIR 79(9) 829-832.pdf329.64 kBAdobe PDFView/Open


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