Pathogen-driven Infectious Disease Recognition and Classification: An In-depth Analysis using Machine Learning and Deep Learning Methods

The human body is a complex structure with well-organized system and consists of unique cells working together to accomplish a specific task for the sustainability of Life. However, there are times when the body is attacked by microorganisms or pathogens, that upset the viable working of the entire framework. These microbes bring about an infection and a condition in the body that makes it hard for the infected cell or organ system to work, this condition is known as an Infectious disease. Pathogens are the main cause for the transmission of infections. Pathogens such as Bacteria, Fungi, Parasites and the dangerous Viruses. The transmission method may vary from one another, some diseases like the flu being transmitted directly from person to person, while others depend on an intermediate vector. This study focuses on pathogen-based disease detection, employing automated techniques for infectious disease detection, classification, and identification of the responsible pathogen. Keras transfer learning models are utilized, considering both open-source public platform datasets and real-time images from hospitals. The primary aim is to investigate and assess various deep transfer learning models to determine the most suitable model for an infectious disease dataset. Various deep transfer learning models are utilized to attain the best classification accuracy by taking x-ray images as datasets, with EfficientNetB1 and EfficientNetB2 demonstrating superior accuracy compared to other models in diagnosing chest infections.