| Resumo: | Accurate quantitative analysis of microorganisms is recognized as an essential tool for gauging safety and quality in a wide range of fields. The enumeration processes of viable microorganisms via traditional culturing techniques are methodically convenient and cost-effective, conferring high applicability worldwide. However, manual counting can be time-consuming, laborious and imprecise. Furthermore, particular pathologies require an urgent and accurate response for the therapy to be effective. To reduce time limitations and perhaps discrepancies, this work introduces an intelligent image processing software capable of automatically quantifying the number of Colony Forming Units (CFUs) in Petri-plates. This rapid enumeration enables the technician to provide an expeditious assessment of the microbial load. Moreover, an auxiliary system is able to differentiate among colony images of Echerichia coli, Pseudomonas aeruginosa and Staphylococcus aureus via Machine Learning, based on a Convolutional Neural Network in a process of cross-validation. For testing and validation of the system, the three bacterial groups were cultured, and a significant labeled database was created, exercising suited microbiological laboratory methodologies and subsequent image acquisition. The system demonstrated acceptable accuracy measures; the mean values of precision, recall and F-measure were 95%, 95% and 0.95, for E. coli, 91%, 91% and 0.90 for P. aeruginosa, and 84%, 86% and 0.85 for S. aureus. The adopted deep learning approach accomplished satisfactory results, manifesting 90.31% of accuracy. Ultimately, evidence related to the time-saving potential of the system was achieved; the time spent on the quantification of plates with a high number of colonies might be reduced to a half and occasionally to a third.
|
|---|