Detection of Knee Osteoarthritis grade using Convolutional Neural Networks

Abstract

Knee OsteoArthritis (KOA) is a disease that affects a person’s quality of life. Early detection and monitoring of KOA progression is essential for effective therapy and quick recovery. A survey of the recent literatures indicates that deep learning methods can effectively assess KOA severity with improved accuracy and efficiency. Convolutional Neural Networks (CNN) help us to classify the levels of severity of Knee Osteoarthritis. The present study proposes a deep learning method in classification of osteoarthritis using Convolution Neural Networks. The study focuses on predicting the grades of input images with KL grades of Knee osteoarthritis. The study employs convolutional neural networks with the Rectified Linear Unit function (ReLU), activation function and Adam optimization algorithm to achieve high performance. The study evaluates 10 performance measures and the results indicate an improvement in performance measures when compared with existing techniques.