Tuberculosis and Pneumonia Detection Using CNN

Artificial intelligence, particularly machine learning, is revolutionizing numerous fields by either supplementing or replacing human efforts, leading to enhanced efficiency and autonomy in systems. Healthcare stands as a notable domain ripe for collaboration with AI and machine learning, offering smoother and more efficient operations. In the context of the modern era, characterized by a scarcity of quality radiologists, the demand for AI-driven solutions in chest X-ray-based disease detection has become increasingly imperative. The subject of this paper is the classification of two major chest diseases, Pneumonia and Tuberculosis, through the implementation of advanced neural network architectures, specifically VGG19 and Convolutional Neural Network (CNN). The system provides diagnostic opinions to users, aiding medical professionals in making prompt and informed decisions about the presence of diseases. In comparison to prior research, this proposed model showcases the capability to detect two types of abnormalities, accurately discerning whether an X-ray is normal or exhibits abnormalities associated with pneumonia and tuberculosis. The VGG19-based CNN achieves remarkable accuracy of 96.87% for Pneumonia and 99.52% for Tuberculosis, surpassing previous models. This advancement underscores the potential of leveraging state-of-the-art neural network architectures for precise and efficient disease classification, addressing the critical need for accurate diagnostic tools in the medical field.