Fig. 1

Schematic figure of the suite of models in AMP utilized in this study. a The plaque segmentation model accepts T2WI sagittal sequences as input and returns the segmentation masks for four organs and the plaques. b The adhesion classification model calculates the radiomics shape features of the organ segmentation results obtained by the plaque segmentation model, employs LightGBM to predict the presence of organ adhesions. c The ovarian cyst segmentation model utilizes TIWI axial sequences as input and generates the segmentation mask for ovarian cysts.