Identification of Malaria Disease Using Machine Learning Models

Malaria, caused by Plasmodium parasites in the bloodstream spread by infected mosquitoes, is a highly severe and sometimes deadly disease. Image analysis and machine learning can enhance diagnosis by quantifying parasitemia on blood slides. The building of an autonomous, accurate, and effective model can significantly reduce the need for trained laborers. This article discusses computer-assisted approaches for finding malaria parasites in blood smear images. These procedures consist of obtaining the dataset, preprocessing the images, segmenting the red blood cells, extracting and choosing features, and classifying the images. The approach is based on well-known Convolutional neural network (CNN) models of Plasmodium parasites and erythrocytes. The trained CNN and VGG-19 are given images of infected and uninfected erythrocytes from the same dataset. VGG 19 gives 96% detection accuracy where CNN achieves 94%.