Classification of Osteo-Arthritis with the Help of Deep Learning and Transfer Learning

Abstract

Osteoarthritis is a degenerative disorder of the joints that negatively impacts the cartilage, bones, and tissues around them. The afflicted joints may experience discomfort, stiffness, and a reduction in range of motion due to osteoarthritis. Additionally, early detection of arthritis can lessen the risk of complications and help in preventing further harm to the affected joints of body. For detection of osteoarthritis in early stage, a model has been proposed in this study which is trained with Knee Osteoarthritis Severity Grading dataset which contains various X-Ray images of joints for healthy, moderate and severe category. This dataset is available on Kaggle. For a better results for accuracy this model was optimized with the help of Adam and Adamax optimizers. Separate values for accuracy, precision, recall and f-1 score has been calculated for both the optimizers. By overall comparison It can be concluded that Adam performs better on this proposed model. Overall accuracy with Adam optimizer is 93.84% and 93.24% for Adamax. Models precision value while using Adam optimizer is 87.80%, recall value is 85.70% and f-1 score is 0.862. Model has performed very nicely with only 102 errors from 1656 test conducted for accuracy of 93.84%. This model can be further used for classification of images for different bone health issues.