Breakthrough of Clinical Candida Cultures Identification Using the Analysis of Volatile Organic Compounds and Artificial Intelligence Methods

Infections triggered by fungi of the genus
Candida
are widely known, although the high incidence and mortality factors are still unclear. The classic methods of identifying
Candida
species are subject to errors, requiring new techniques with faster and more accurate performance. We present a study for identifying fungi species by analyzing volatile organic compounds of cultures acquired and interpreted using Electronic Nose and Artificial Intelligence methods. The proposed approach contributes to establishing an agile and appropriate treatment, reducing the complications of the disease and the number of deaths. We perform experiments with three species of
Candida
obtaining accuracy above 90% in the fungi identification. Therefore, future works are encouraged to deal with more types of fungi to help create a new identification methodology faster and more reliable using artificial intelligence methods.