a deep learning model for joint detection in radiographic images of rheumatoid arthritis

Rheumatoid arthritis (ra) is an autoimmune disease that leads to joint deformity and permanent disability. in this research, a deep learning method is proposed for the automatic detection of joints on the radiographic images of the hands of ra patients. the proposed byolo model uses the backtracking search algorithm (bsa) to tune the optimal values for the hyper-parameters of the yolo. five-fold cross-validation is used to evaluate the performance of system. a dataset containing hand images of 200 patients has been used to evaluate the proposed model. the results obtained based on performance criteria such as accuracy and error rate show that the proposed model is promising in diagnosing joints in radiographic images.