Quadratic Program Optimization using Support Vector Machine for CT Brain Image Classification
In this paper, an efficient Computer Tomography (CT) image classification using Support Vector Machine (SVM) with optimized quadratic programming methodology is proposed. Due to manual interpretation of brain images based on visual examination by radiologist/physician that cause incorrect diagnosis, when a large number of CT images are analyzed. To avoid the human error, an automated optimized classification system is proposed for abnormal CT image identification. This is an automated system for content based image retrieval with better classifier accuracy and prediction time. SVM classifier can accurately train up the datas as normal and abnormal brains interpreted manually by the user. The system can retrieve more number of images present in the query data base. The proposed classifier is analyzed with existing Sequential Minimal Optimization (SMO) and K Nearest Neighbour classifier KNN). From the experimental analysis, the proposed classifier outperforms all other classifier taken for examination.
Keywords: CT Brain image classification, Quadratic programming, Linear SVM, SMO, KNN.
Download Full-Text
ABOUT THE AUTHOR
J.Umamaheswari
FullTime, Research Scholar, ComputerScience
J.Umamaheswari
FullTime, Research Scholar, ComputerScience