Publication

Abstract : Early diagnosis and treatment of pneumonia are critical, and they have a substantial influence on patient outcomes. We provide a thorough comparison of pneumonia diagnosis models using both conventional X-ray imaging and modern CT scans in this research study. In this work, we examine how well deep learning (DL) models do in accurately identifying pneumonia using both imaging modalities. Furthermore, we investigate the application of Explainable Artificial Intelligence (XAI) methodologies to better understand these models’ decision-making procedures and establish confidence in their projections. The research utilizes a mix of dataset that includes CT and X-ray pictures of pneumonia patients that were obtained from several web sources. To determine how well deep learning models for each imaging modality identify pneumonia, we train and test them. We employ XAI approaches to clarify decision-making processes within the models, offering insights into the characteristics that influence the classification results. Our results provide insight into the relative effectiveness of pneumonia detection models in CT and X-ray scans through thorough testing and assessment. Additionally, the combination of XAI and Pneumonia detection model approach shows encouraging progress in improving the robustness and interpretability of models, opening the door for more dependable and beneficial diagnostic systems in the field of medical imaging.