A Comparative Study of Artificial Intelligence and eXplainable AI Techniques for Pulmonary Disease Detection and Its Severity Classification

As respiratory diseases like tuberculosis, pneumonia, COVID-19, and cancer become more common, the need for accurate and efficient diagnostic tools is growing. This paper provides an overview of the use of artificial intelligence (AI) and explainable AI techniques for pulmonary disease detection and severity classification. Using AI, especially deep learning (DL) algorithms, has shown promise in automating the process of detecting pulmonary diseases from chest radiographs. This study looks at different AI methodologies, such as deep learning architectures and preprocessing techniques, and uses explainable AI (XAI) to improve pulmonary disease detection and severity classification mechanisms. It also addresses the restrictions and challenges associated with AI-based disease detection, offering insights into possible future directions. To group cutting-edge techniques according to image types, DL algorithms, XAI techniques, and targeted pulmonary diseases, a taxonomy is put out. By helping researchers organize their work and contributions, this taxonomy eventually improves the efficiency of AI-assisted pulmonary disease detection and severity classification systems.