Performance of SVM and ANFIS for Classification of Malaria Parasite and Its Life-Cycle-Stages in Blood Smear

Abstract

A method to classify Plasmodium malaria disease along with its life stage is presented. The geometry and texture features are used as Plasmodium features for classification. The geometry features are area and perimeters. The texture features are computed from GLCM matrices. The support vector machine (SVM) classifier is employed for classifying the Plasmodium and its life stage into 12 classes. Experiments were conducted using 600 images of blood samples. The SVM with RBF kernel yields an accuracy of 99.1%, while the ANFIS gives an accuracy of 88.5%.