Discovering Knee Osteoarthritis Using CNN Enhanced with AlexNet

Pain and a decline in gait function are symptoms of knee osteoarthritis (KOA). With the use of machine learning methods, developed regression models to identify KOA-related gait variables according to outcome measurements (PROMS) that the patient reports. 375 people with a range of KOA scores took part in the study. Making use of the McMaster Universities and Western Ontario Osteoarthritis Index (WOMAC), the severity of KOA was assessed. The severity of the diseases was divided into three groups based on WOMAC scores. From the gait data, 1087 characteristics in total were retrieved. After doing an ANOVA and student's t-test, the machine learning system selected only significant features. Each of the three subscales of the WOMAC: stiffness, discomfort, and physical function has three categories. To ascertain which chosen traits were substantially correlated with the subscales, an ANOVA was conducted. To estimate the patient's WOMAC values, a random forest regression as well as linear regression models were utilized. Based on the outcomes of the student t-test and ANOVA, 43 characteristics were chosen. Twelve characteristics from the hip, one from the pelvis, seventeen from the foot, three from the spatiotemporal characteristics, nine from the knee, and one from the ankle were chosen from each joint. This research proposed a novel approach utilizing Convolutional Neural Networks (CNNs) with the AlexNet architecture to enhance the diagnostic accuracy and severity assessment of knee osteoarthritis. The study leverages a large dataset of knee radiographic images, encompassing diverse cases of OA progression. The AlexNet architecture, renowned for its deep learning capabilities in image classification tasks, is employed to automatically extract hierarchical features from the input radiographs. The CNN model is trained to discern subtle patterns and nuanced information indicative of OA-related structural changes, enabling it to differentiate between varying stages of the disease.