Knee Osteoarthritis Severity Prediction Through Medical Image Analysis Using Deep Learning Architectures

Like millions of individuals knee osteoarthritis (OA) is a common degenerative joint condition that particularly affects individuals, over the age of 50. The primary cause of knee osteoarthritis is the breakdown of cartilage in the joints leading to friction swelling and stiffness in the knees. It’s important for people to regularly assess their knee health as 40% of individuals will experience osteoarthritis at some point in their lives. Our goal is to accurately identify the five stages of knee osteoarthritis through the use of x ray images and a deep learning model. In this report we examine CNN designs, such as convolution based networks, like VGG, ALEXNET, RESNET and LENET. We found that the LENET architecture achieved an accuracy of 89.0% in classifying the results. The prognosis will be updated through a web application where clinicians can view x ray images and provide patients with advice based on their stage of osteoarthritis. As a result, patients can log in. Check their knee condition as well as receive expert recommendations.