Development and validation of a CNN-based model for Malaria Cell Detection

Abstract

Malaria remains a great threat to the African continent and the world at large, and accurate Plasmodium detection and treatment are the most effective means of preventing the progression of mild malaria into severe disease and into death finally. However, manual operations and diagnosis face the challenge of individual and geographic heterogeneity. In order to solve this problem, this study proposed a Convolutional Neural Networks (CNN) based model for malaria cell detection, and the model can automatically recognize cell appearance and distinguish infected cells from uninfected cells, leading to an accurate and effective cell classification. The research results showed that CNN has better performance in malaria cell recognition compared with the traditional integrated learning model, and the optimized CNN algorithm achieved an accuracy of 98.29% compared with the average accuracy of integrated learning algorithms of 80.29%, which can effectively solve the problem of wrong and missed detection of in malaria cell detection