CNN Classification of Malaria Parasites in Digital Microscope Images Using Python on Raspberry Pi

Abstract

Malaria disease occurs because the plasmodium parasite infects human red blood cells spread by female Anopheles mosquitoes and then causes health problems such as blood deficiency and even death. The gold standard of malaria diagnosis is to use laboratory microscopy examination of the patient's red blood cell samples to distinguish between microscope images of parasitic and non-parasitic red blood cells. However, diagnosing malaria through microscope observation is time-consuming, subjective, and tiring for health workers. So, a malaria classification system was designed using the Convolutional Neural Network (CNN) method to distinguish parasitic and non-parasitic red blood cell images. The CNN model is trained using training data and also tested using test data. Then, the CNN training model is embedded on a Raspberry Pi equipped with a Graphical User Interface to facilitate observer interaction through the LCD screen on this digital and portable microscope. The CNN classification rate achieved an accuracy value of 97.88% using the database image and 98.76% using the digital microscope acquisition image. The CNN classification system of malaria parasites designed on a Raspberry Pi-based digital and portable microscope is expected to improve the diagnosis of malaria and reduce the infection rate of malaria patients, especially in various remote areas in Indonesia.