Malaria Parasite Detection in Thick Blood Smears using Deep Learning

Abstract

Malaria parasitized detection is very important to detect as there are so many deaths due to false detection of malaria in medical reports. So analysis has gained a lot of attention in recent years. Detection of malaria is important as fast as possible because detecting malaria is difficult in blood smears. Our idea is to build a transfer learning model and detect the thick blood smears whether the presence of malaria parasites in a drop of blood. The data consists of 5000 each infected and uninfected data obtained from the NIH website. In this paper, I propose to use three different types of neural networks for the performance evaluation of the malaria data by transfer learning using CNN, VGG19, and fine-tuned VGG19. Transfer learning model performed well among various other models by achieving a precision of 98 percent and an f-1 score of 96 percent.