A Proposed Mala-Net CNN Algorithm for Malaria Early Detection Based on Thick Blood Smear
Zul Indra, Yessi Jusman, Elfizar, Evfi Mahdiyah, Roni Salambue, Doni Winarso · 2024
Malaria has spread worldwide since the early 20th century and causes nearly half a million deaths each year. Malaria is actually not a dangerous disease and can be easily cured if the treatment can be carried out effectively. Therefore, efforts to identify the presence of the plasmodium parasite early will greatly help reduce the death rate caused by this disease. However, the fact is that this disease often does not get serious attention. Generally, this disease is only treated in the critical phase because it is considered just a common flu. This research is intended to produce a computer-aided disease diagnosis (CAD) system enriched with CNN algorithms to assist in early malaria diagnosis. This CAD system is expected to provide an effective and reliable malaria diagnosis, and avoid the limitations of manual diagnosis. In order to develop the CAD applications, this study proposed a novel CNN architecture called as Mala-Net by utilizing red thick blood dataset. In conclusion, this study successfully surpassed previous studies used as benchmarks with an accuracy value above 97%.