Implementation of Malaria Parasite Detection and Species Classification Using Dilated Convolutional Neural Network

Abstract

Malaria is an infectious disease caused by a bite of the Anopheles Mosquito which has caused a lot of death. Treating malaria infected patient involves diagnosis using microscopy to ascertain its presence in the blood. This microscopy process takes time, need for a laboratory expert to read the results obtained and changes involved in the morphology of the parasites life cycle stages. To get reliable and accurate diagnosis, machine learning has been used to automate this process; still there is issue of enough datasets and high computational time. This paper presents an implemented Dilated Convolution Neural Network for malaria parasites detection and species classification using blood smear images. A direct classification was carried out to detect infected and uninfected malaria parasites and subsequently species classification was carried out whereby 3 convolutional layers was deployed, Convolution2D was for used for convolution operation and dilation rate of 2 was fixed. The model was trained with public available dataset for comparison with other existing models. This model shows overall performance accuracy as criteria of 99.9% for parasite detection and 88.9% for the species classification. We were able to achieve better results in comparison with other experiment and most importantly we deployed it to classify four species of malaria parasites; falciparum, ovale, malarie and vivax.