Monitoring Bone Healing: Integrating RF Sensing with AI

This study develops a refined machine learning model based on Radio Frequency (RF) sensing for monitoring bone fractures, particularly the femur. Utilising a dataset enhanced with variations in bone and soft tissue properties and augmented with multiple sensor inputs, the model aims to provide a more frequent and non-invasive assessment of the fracture healing process compared to conventional imaging methods. Leveraging data from six sensors, our approach achieves notable accuracy in detecting ‘no fracture’ states and classifying different hematoma sizes. The results indicate a significant potential for this technology to inform and improve personalised patient care plans by offering continuous monitoring without radiation exposure.