Dementia MRI Classification Using Hybrid Dragonfly Based Support Vector Machine

Automated medical image classification can help the clinician during diagnosis and treatment of disease to a greater extent. The objective of this research work is to analyze the brain MRI images for classifying the subjects into two classes: Dementia & non-dementia subjects. In this study, MRI images of 65 non-dementia and 52 dementia subjects downloaded from Open Access Series of Imaging Studies website is analyzed using Matlab software. Numbers of literatures are reported regarding the usage of Dragonfly algorithm in parameters optimization for support vector machine. But it is inventive to use the Dragonfly algorithm for transforming the input data into the data more appropriate for successful support vector machine implementation. The popular Principal Component Analysis based Support vector machine is used as benchmark technique for performance comparison. It is observed that individual Support Vector Machine is able to provide accuracy of 67.57% while hybrid dragonfly based Support Vector Machine provides 81.08% accuracy and Principal Component Analysis based Support vector machine provides 70.27% accuracy.