Intelligent prediction method for recurrence of hepatolithiasis based on multi-source heterogeneous data fusion

The invention provides a multi-source heterogeneous data fusion-based intelligent prediction method for recurrence of hepatic duct calculus, and the method comprises the following steps: S1, obtaining a CT two-dimensional image and an MRI two-dimensional image of the abdomen of a hepatic duct calculus patient, and carrying out the image reconstruction of the CT two-dimensional image and the MRI two-dimensional image through a diffusion model, image depth features are extracted from the reconstructed CT three-dimensional image and the reconstructed MRI three-dimensional image; preoperative blood test indexes, hepatic duct stenosis/expansion degree data and hepatic duct disease occurrence time sequences of the hepatic duct calculus patients are obtained, and fusion features of clinical data are extracted and obtained; and S2, carrying out feature fusion on the image depth features and the fusion features of the clinical data, and then inputting the image depth features and the fusion features into a pre-trained