The system brings efficiency to the analysis of microscopic images of blood samples by automatically identifying the size and shape of infected cells, as well as classifying with valuable precision the species and stages of development of the parasites responsible for malaria.

According to Anto Satriyo Nugroho, director of BRIN’s AI and Cybersecurity Research Center, the model has been trained with more than 1,300 images and achieves an accuracy of 80,6% in detecting the disease.

The technology is able to identify the four main species of malaria parasites present in Indonesia, the plasmodium: falciparum, vivax, malariae and ovale.

One of the main challenges of the project is the morphological variability of the parasites throughout their life cycle, which requires a high degree of accuracy in AI training, Nugroho explained.

The progress of this tool is particularly relevant for regions such as Papua, where most of the country’s malaria cases are concentrated and where a shortage of trained personnel makes traditional diagnosis difficult.

jrr/arm/oda/msm