DNN-based Knee Osteoarthritis Disease Detection using X-Rays

A degenerative condition of the knee joint, knee osteoarthritis (KOA), is brought on by deterioration of the ligaments that unite the tibia bones and femur. Clinically, KOA is categorized into four classes according to the degree of degeneration of the connective tissue that connects these two bones, which limits range of motion. It is highly subjective and challenging to determine this interosseous region with a front X-ray examination of the knee. In the automatic separation and categorization of different knee tissues and structures, including soft tissue, bone, and cartilage, DNN has demonstrated promising results. This is important since it helps diagnose the degree and course of osteoarthritis. This research presents a deep learning method utilizing knee radiographs to automatically identify and rate the severity of KOA. Our structure of deep neural networks is capable of precisely identifying knee X-rays as either unaffected or KOA-affected and estimating the degree of KOA. Our deep neural network model was created utilizing a large collection of knee joint X-rays that had been expertly annotated by radiologists.