Transfer Learning Based Intelligent Diagnosis Support System for Knee Osteoarthritis

— Recent studies have determined that osteoarthritis has become one of the most common diseases of the knee joint. As an irreversible disease, with the gradual wear and tear of articular cartilage, its condition will gradually deteriorate, which will bring inconvenience to patients' lives. As the main criterion of osteoarthritis, knee x-rays usually require professional judgment from doctors with rich experience. However, a large number of x-rays and small differences between different stages often cause a heavy burden on doctors and make mistakes in judgment. With the rapid development in recent years, machine learning technology or artificial intelligence (AI) has become a common solution to medical problems. In this research, a transfer learning based medical diagnosis support system was built to reduce the burden on physicians. Furthermore, a friendly user interface was designed for medical staff to efficiently access the functionality of the proposed system. After comparing different algorithms with different frameworks, a transfer learning model with 96.98% accuracy built into the system