A Comparative Analysis of Deep Learning Models for Malaria Plasmodium Classification

Abstract

Malaria, caused by the Plasmodium parasite and transmitted by Anopheles mosquitoes, poses a significant health threat in sub-Saharan Africa, especially in resource-limited settings. This study aims to address this by analyzing three advanced digital and biomedical technologies in malaria detection and species identification. We used a dataset of 502 Giemsa-stained thick blood smear images from Rwanda to evaluate the efficacy of these models in detecting and classifying four Plasmodium species: P. falciparum, P. malariae, P. ovale, and P. vivax. The results showed that YOLOv5 demonstrated superior overall performance, par-ticularly in multi -class detection scenarios, suggesting its potential for early and accurate diagnosis in real-world applications. Mask R-CNN showed the best performance in detecting P. falciparum, while Faster R-CNN exhibited consistent performance across all species. This research contributes to the development of computer-aided diagnostic tools tailored for regions with limited resources, aiming to enhance malaria control strategies in developing areas.