Published Paper Details:

Abstract Rheumatoid Arthritis (RA) is an autoimmune disorder, which can lead to joint deformation and which can lead to permanent disability that cannot be treated in a timely manner. Early diagnosis is thus imperative for effective treatment. In this paper, we present an end-to-end RA detection system with the help of Convolutional Neural Networks (CNN) for hand X-ray images. The model is trained on a balanced 200-image dataset (100 RA, 100 Normal) with extensive pre-processing and augmentation techniques which is used to enhance performance in the lack of large datasets. The CNN architecture was here we used is chosen for its ability to extract hierarchical features from medical images without manual feature engineering. The trained model was deployed using a Flask-based web application, with real-time prediction supported by confidence scores and easy-to-use interaction.This Experiment gives the results that show high reliability, with test cases reporting 96.24% confidence for Normal x-ray image and 92.27% for RA-positive cases.There is another approach where the numerical clinical dataset is used which contain details such as age, gender, ESR, CRP, RF, and Anti-CCP levels. Here the Deep Neural Network (DNN) was trained after preprocessing and scaling the data, demonstrating high predictive performance compared to traditional Machine Learning models. This model was implemented using a Flask-based web application, which will work with real-time prediction and confidence scores and simple interaction. The system can be potentially integrated into clinical workflows, especially in regions where specialist availability is scarce.