Lung cancer risk prediction method based on cross-modal knowledge distillation

The invention belongs to the field of medical health big data, and particularly relates to a lung cancer risk prediction method based on cross-modal knowledge distillation, and the method comprises the steps: selecting a trained I3D model based on CT image input as a teacher model; constructing a student model by adopting a multi-layer perceptron model; inputting the CT image data set into a teacher model to obtain a first intermediate feature vector set; taking the structured data as student input to train a student model, and defining a feature extracted from the structured data in a student model training process as a second intermediate feature; calculating teacher model space distribution and student model space distribution based on the first intermediate feature vector set and the second intermediate feature vector set; calculating divergence loss between the teacher model space distribution and the student model space distribution by adopting KL divergence; calculating the weighted sum of the divergen