Rapid bacterial colony classification using deep learning

Abstract

Bacterial colonies infection is one of the causes of bloodstream disease, and it can be a fatality. Therefore, medical diagnoses require fast identification and classification of organisms. Artificial Intelligence with deep learning (DL) can now be developed as a rapid bacterial classification. The research aims to combine deep learning and support vector machines (SVM). The ResNet-101 model of the DL algorithm extracted the image’s features using transfer learning then classified by the SVM classifier. According to the experimental results, this model had 99.61% accuracy, 99.58% recall, 99.58% precision, and 99.97% specificity. The technique presented might enhance clinical decision-making.