Application of MobileNetV2 Architecture to Classification of Knee Osteoarthritis Based on X-ray Images

Knee osteoarthritis is a health problem in the knee that is experienced by most people and causes difficulty in moving normally. Knee osteoarthritis can be caused by several factors such as age, active smoking, weight, and family history of suffering. However, this is often ignored and taken for granted by some people, although it can be fatal, such as paralysis, owing to decreased cartilage volume. Based on existing problems, a system that can classify knee osteoarthritis is required. Utilizing a convolutional neural network (CNN) to classify X-ray pictures allows us to take use of advances in science in the field of artificial intelligence. Therefore, this study suggests using a convolutional neural network design to classify osteoarthritis in knees based on Xray images. Four research techniques were used in this study: classification of X-ray pictures of the knee, preprocessing, CNN with MobileNetV2 architecture, and X-ray image datasets of the knee. According to research findings, further advancement is still required when classifying knee images in order to detect osteoarthritis disease. Several architectures, including VGG, Dense Net, Inception, MobileNetV2, and CNN, were used, and the accuracy rate for the MobileNetV2 architecture was 94%, with a loss of 24%, and a computation time of 6 minutes.