Malaria Parasite Classification Employing Chan–Vese Algorithm and SVM for Healthcare

Abstract

Malaria a contagious disease that spreads through Anopheles mosquito bite, it’s deathly infection which if not treated in time can take lives and become an epidemic. According to WHO, it is already one of the major reasons of death in tropical and sub-tropical regions across the globe. Early detection and diagnosis of malaria can help to prevent lot of casualties; however, gold test till date is by microscopy. Different species of plasmodium parasite have evolved over generations, five main that can be listed are: P. falciparum, P. vivax, P. ovale, P. malariae, and P. knowlesi, these parasites affect human adversely. Deadliest among these is P. falciparum which contributes to majority of deaths and has also developed resistances to various drugs employed to prevent the infection. Accurate evaluation of parasite density and, hence, identification of infection till date are done by an expert; which is an exhaustive process and lack accuracy. Paper presents an automated algorithm based on SVM for accurate and early determination of parasite density and infection in red blood cells, the algorithm was designed for automated detection and was tested on sample of 100 images obtained from online UCI repository. An efficiency of 91% percent was achieved on sample infected images utilizing SVM classifier.