Patent attributes
Embodiments of the present invention relate to methods of predicting fracture risk, which improve fracture risk prediction by developing a bone radiomics score model based on machine learning. As an embodiment of the present invention, the method of predicting the fracture risk is configured to perform the steps of designing a development set, processing bone images for a plurality of subjects included in the development set, extracting texture features from the bone images, selecting optimal texture features required to predict the fracture risk from the extracted texture features, performing machine learning for the optimal texture features using a training set of the development set, and designing a bone radiomics score model to predict the fracture risk.