Advanced Knee Osteoarthritis Detection using Deep Learning

This present research investigates the use of machine learning and deep learning models for the detection of knee osteoarthritis using the X-Ray images in depth. knee osteoarthritis is one of the most common joint disorders which affects numerous percentages of people across the world. And its detection of it at an early stage is very critical for effective management. This paper presents valuable dataset of X-ray images with a dual focus on detection of knee joint and KellgrenLawrence grading [kl]. The methodology which is proposed in this paper includes basic steps such as data loading and preprocessing. It also includes feature extraction and utilization of models of both machine and deep learning. In the phase of machine learning models which are used are Decision tree, KNN, MLP, Random Forest and ensemble methods. While in deep learning models used for evaluation are Nasnetmobile, VGG19, Darknet53, Darknet19 and Googlenet. The output of all the above-mentioned models are compared and evaluated based on four most important metrics of performance. The final output of the research shows that Random Forest and the Ensemble methods have outperformed in the phase of machine learning and when it comes to deep learning Googlenet and Darknet53 have done well in comparison. This study helps in the early detection of the knee osteoarthritis and also helps in management of it by giving valuable insights to medical practitioners and researchers.