KL Grading-Based Knee Abnormality Classification Using Multi-Modal Deep Learning: KOMMCF – A Fusion of X-Ray Imaging and Clinical Biomarkers

Abstract

Injuries in the knee or any wear and tear of the joint causes knee abnormalities. Here, we propose a light weight novel model "Knee Osteoarthritis Multi-Modal Classification Framework (KOMMCF)" aimed at early detection of knee abnormalities by leveraging the KL grading system. We have trained our model by using X-ray images ranging from [grade 0 - 4 normal to severe] and 14 clinical biomarkers and evaluated the model using both validation and test datasets. We have achieved [89% - validation and 88% - test] overall accuracy and classified model in 3 classes Normal, Early Stage (doubtful/mild), and Advanced Stage(moderate/severe) by using Kaggle dataset. We have also employed CAM to visualize areas of knee X-ray images that contributed most in the model's predictions. This work demonstrates the power of combining both imaging and clinical data in an AI-driven approach to knee abnormality classification, potentially offering a valuable tool for early diagnosis and treatment planning.