Malaria Detection Using Deep Learning: An In-Depth Comparison of Techniques

The Plasmodium parasite, which is accountable for malaria, is a lethal disease that infests individuals once female Anopheles mosquitoes bite them. It persists as one of the most prevalent and deadly diseases in tropical and subtropical regions. Prompt Identification of malaria is necessary for successful therapeutics and deterrence of acute repercussions. Microscopic is the standard of excellence for diagnosing malaria. However, the result of microscopic analysis takes time and depends on factors such as quality of equipment, lighting conditions, and the expertise of parasitologists. Coupled with traditional detection methods, experts have explored microscopic picture examination rooted in deep learning models to recognize Plasmodium parasites. This study proposes a comparative analysis of the prediction of malaria parasites and healthy cells using different deep learning models and open-access datasets. Several measures are employed to evaluate the efficacy of various architectures. The outcomes show that ResNet-50 outperformed delivering an accuracy rate of 96.88%. The findings allude that neural network architectures can effectively diagnose malaria while identifying the optimal model is important in order to achieve high accuracy.