ARTHRO—Knee Osteoarthritis Detection Using Deep Learning

Knee osteoarthritis (KOA) is a frequently arising orthopaedic condition that impacts a larger number of the population worldwide, causing significant pain and disability. Early detection of KOA is critical for timely intervention and improving patient outcomes. In this research, we develop a KOA prediction model using X-ray images and the Kellgren-Lawrence (KL) scale to predict the presence of KOA. Medically, photos in the form of knee X-ray images were collected and labelled by radiologists based on the KL scale, which ranges from 0 (no KOA) to 4 (severe KOA). Convolutional neural network (CNN) modelling is done to interpret X-ray pictures and forecast the KL score. This developed model attains a great degree of accuracy in predicting the KL score, with an overall accuracy of 81%. The model also demonstrates high sensitivity and specificity in detecting different levels of KOA. We conducted an evaluation of the model utilizing performance indicators like accuracy, model loss, F1 score, recall and precision. The study results suggest that the developed KOA prediction model can accurately predict the presence of KOA on the KL scale using X-ray images. The model has the potential to assist healthcare providers in early detection of KOA. This project demonstrates the potential of deep learning algorithms, such as CNN, to develop KOA prediction models using X-ray images and the KL scale.