A BPNN for the identification of Rheumatoid arthritis

RA is a dangerous and chronic disease that should be analyzed and detected in its early stages. Thus, the aim of this thesis is to develop a new approach for the identification of rheumatoid arthritis through knee image processing techniques and neural network classifier. Thus, any supplied knee image must be classified either normal or abnormal. The proposed system uses knees images obtained from a created database comprised of 300 images for training and 300 images for testing phase. The used image processing techniques facilitate the diagnosis of that disease by analyzing and pointing out the rheumatoid arthritis signs and symptoms through extracting the useful and needed features or patterns such as the distance between the femoral and tibial bones, joint damage, bone widening and stiffening. Moreover, the developed system helps the doctors to accurately classify the RA knee X-ray images since it is designed to stimulate the human visual inspection that is based on visualizing some related features and signs of RA