Deep Convolutional Neural Network Classifier for Effective Knee Osteoarthritis Classification

Abstract

Millions of people are affected by the disease Knee Osteoarthritis, and the prevalence of the condition is steadily increasing. Knee osteoarthritis has a significant impact on people's lives by generating increased worry, mental health disorders, and physical problems. Early detection of knee osteoarthritis is critical for decreasing disease consequences, and numerous studies are being conducted to classify knee osteoarthritis. In this study, the deep CNN classifier is used to classify knee osteoarthritis, which effectively extracts the features required for disease classification more efficiently. The preprocessing of the data, which is done in three processes such as Circular Fourier Transform, Multivariate Linear Function, and Histogram Equalization, is particularly important in this research since it aids in obtaining more efficient information about the image. The deep CNN classifier's weights and bias deliver better and desired classification results while spending less time and storage. The proposed deep CNN classifier attained the Accuracy of 94.244%, F1 measure of 94.059%, Precision of 94.059%, Recall of 93.586%.