A Hybrid Approach for Classifying Drug-Resistant Microscopic Species

Like all living things, bacteria play a crucial role in a variety of industries, including agriculture, food processing, medicine, and others. Only a small percentage of them are helpful to people. However, most of them are harmful to people and can lead to fatal illnesses. Hence timely identification and taking necessary action is very important. Therefore, it's important to identify them quickly and take the appropriate steps to stop them. The first step in this would be the identification and classification of bacteria. Traditional or manual methods for classification include gram staining, motility testing, biochemical testing, and so on. But these methods are time-consuming and also require human expertise along with long training time to train the operators. Additionally, with the advancements in technology, a system for the automatic classification of bacteria can be developed using machine learning algorithms. Therefore, in this paper, we aim in developing a system that uses machine learning techniques to create a system for automatically classifying bacteria, cutting down on the time required for the human procedure. As CNN is proven to be highly efficient for working with images, we are employing CNN for image processing. But due to the requirement of larger number of images for CNN training we are using a hybrid approach. Random Forest has been employed along with CNN for efficient classification. The system proved to be efficient with an accuracy of 98%. The inclusion of Artificial Intelligence is greatly assisting microbiologists, the food industry, cosmetics, pharmaceuticals, and many more.