Detection and Classification of Knee Osteoarthritis Using Convolutional Neural Network
M. Janotheepan, Sudeera Wickramarathna, S. Bavathaaranie, M. Yanusha, K. Subahari, K. Thanusan
Abstract
Knee osteoarthritis (KOA) is a prevalent degenerative joint disease that significantly impacts the quality of life for millions of individuals worldwide. Early and accurate detection of KOA is crucial for timely intervention and effective management of the disease. This study propose a deep learning model to enhance the accuracy and cost-effectiveness of KOA detection using X-ray images. Initially this study evaluate the performance of a Convolutional Neural Network (CNN) model for the detection and classification of KOA using X-ray images. The CNN model will be trained on a large dataset of knee X-ray images, annotated by expert radiologists, to learn the patterns and features indicative of KOA. Then analyze the impact of different image preprocessing techniques and hyper-parameters on the performance of the CNN model. By systematically exploring various preprocessing methods, such as image normalization, filtering, and enhancement, along with hyper-parameter tuning. Finally develop a web interface that allows individuals to conveniently upload their knee X-ray images and receive a quick and accurate prediction of having KOA. Furthermore, insights gained from the analysis of image preprocessing techniques and hyper-parameters will contribute to the optimization of KOA detection systems. Overall, this research study aims to advance the field of KOA detection by harnessing the power of deep learning and developing a practical tool for screening and assessment. Especially this study achieves the accountable amount of accuracy as 89%. The proposed model and web interface have the potential to improve healthcare outcomes by enabling timely diagnosis and intervention, ultimately enhancing the management of KOA and improving the quality of life for affected individuals.