Image processing for detection of dengue virus based on WBC classification and decision tree

Abstract

Dengue is a major health problem in tropical and Asia-Pacific regions which typically spreads rapidly in number of infection patients. Knowing that most of the world's population living in risk areas, in order to diagnose and treat the disease, high skilled experts and human resources are needed. However, in some cases human error potentially may occur. Therefore, in this research we developed a model which can diagnose dengue fever disease. This study used blood smear images that were taken under a digital microscope with 400 × magnification specifications by means of image processing techniques such as color transformation, image segmentation, edge detection feature extraction and white blood cells classification. In this study we used white blood cell counting of the role of cell differentiation as a new feature that can classify dengue viral infections of patients via decision tree methods. The results showed that, the white blood cells classification modeling technique of 167 cell images resulted in 92.2% accuracy while dengue classification modeling technique of 264 blood cell images resulted in 72.3% accuracy.