Knee Osteoarthritis Detection and Classification using Ensemble Model of Xception and Inception ResNetV2

Knee Osteoarthritis is one of the predominant types of arthritis, affecting joint tissues and afflicting millions of people globally. Currently, there are no curative medications exist for this condition, so there is a need to focus on early detection to manage disease progression. X-ray imaging serves as a primary method for osteoarthritis prediction, but manual prediction may sometimes be error-prone due to limited radiologist expertise. Recent studies have explored many machine learning systems for efficient osteoarthritis prediction from X-ray images, yet upgraded predictive accuracy is essential for early-stage detection. This study targets knee osteoarthritis detection and classification using Kellgren and Lawrence's grading scheme, utilizing an ensemble model combining Xception and InceptionResNetV2 architectures. The ensemble model excels with a 95.1% accuracy score, contributing to medical image analysis and providing clinicians with a robust tool for early diagnosis and treatment planning in knee osteoarthritis.