Comparative Analysis of CNN, Xception and Inceptionv3 for Classifying Tuberculosis, Pneumonia Normal Chest X-ray Images

In the 21st century, health has become a primary concern for every individual. Despite their overall health awareness, people often spend money on unnecessary medical scans, possibly due to weather changes or other factors. Some diseases or issues can be linked to specific organs or parts of the human body. For instance, a deep learning model can use chest X-ray images to classify conditions such as pneumonia, tuberculosis, or other diseases. This study proposes a solution to diagnose multiple issues using chest radiographic images, employing models like InceptionV3, Xception, and CNN. These models achieve accuracies of 94%, 90%, and 87% respectively. The accuracy can be further enhanced by developing hybrid or ensemble models. This approach aims to optimize medical diagnoses, reducing unnecessary scans and focusing on specific health concerns effectively.