Bacteria play an important role in human life since they influence all aspects of life, from vital processes within the human body to the production of medical drugs and vaccines, as well as food production. As this stage is considered one of the basic stages in the diagnosis process, the Bacilli shape is one of the basic forms of bacteria that microbiologists use medical microscopes to diagnose. The purpose of this Article is to develop a bacilliform diagnosis system that employs a pre-trained ResNet-50 algorithm as a feature extraction layer to train the Faster R-CNN detector model.DIBaS (Digital Images of Bacteria Species) dataset, is a public dataset containing 33 different types of bacteria used in training and validating the system. The proposed system achieved 98.99% mini-batch accuracy and 99.10% validation accuracy.

Deep learning, ResNet-50, Faster R-CNN, Bacteria, Bacillus spp