Development of a Website for Malarial Detection using Deep Learning

Abstract

Malaria is a bacterial disease which is commonly caused by mosquitoes. In India, approximately, 7500 people have died due to malaria. The screening methods for malaria include the analysis of blood samples. This research aims the development a website that is capable of predicting malaria by analyzing the images of blood cells. For this purpose, images of blood cells are collected as a dataset from Kaggle. Both healthy and blood cells infected with malaria are included in this dataset. Next, the dataset's photos are selected for training, testing, and validation. The picked pictures are then scaled down to a specific size. With the help of the Convolutional Neural Network (CNN), a Deep Learning (DL) model is created. The preprocessed photos are then used to train this model. Model validation follows the training. The training and validation results are tallied and examined. Next, the model's accuracy and loss are evaluated. The highest accuracy of the model developed is 9% which was attained during the training. The model also produced the lowest loss value of 13% during the final epoch of the validation process. The model is then tested if the findings are satisfactory. The tested model is then deployed on a website. This website can be used as a pre-screening test for malaria in times when a person cannot reach out to the nearest doctor. This website can also be updated and converted as a software application in the future.