AN AUTOMATED APPROACH for DIAGNOSIS of MALARIA DISEASE

Abstract: Malaria [1] is a dangerous disease, infection starts when an infected Anopheles mosquito injects plasmodium parasites into the skin of the human. However, patients who have been diagnosed with malaria can get early medical and clinical therapies to improve their possibility of surviving. Therefore, it's crucial to find malaria early on. In this research, we used a trained Convolutional Neural Network [6] to tackle the issue of detecting malaria through a web application. The thin blood smear image is given as input into the application, which then outputs the blood cell's class, which can be either benign (milder) or malignant (harmful), depending on the image. The model used in the web application had a validation loss of 0.3536 and an accuracy of 0.9588.

Keywords: Malaria, Thin Blood Smear, Web Application, Convolutional Neural Network (CNN).