关键词: 肝脏模型, 个性化, 高精度, 模型驱动

Abstract: The personalization of 3D liver models and the acquisition of liver models with high precision are two key technologies in virtual surgery systems. A personalized data-driven method aiming at the problems of time efficiency and precision of liver models was proposed. The liver model with high precision from previous research achievements and personal CT datasets were processed respectively, and then some feature points were marked on both the model and the CT datasets. The PCA algorithm was used to get the initial matching, the ICP algorithm was used to get the matching of the feature points, and then an improved ICP algorithm was used to calculate and get the dense correspondence of two datasets based on the matching results of previous matching algorithms. The dense correspondence of the two datasets was used to lead the liver model to perform the RBF deformation to finally acquire the personalized 3D liver models. The results show that the algorithm is time-efficient and can get ideal 3D liver models using this algorithm.

Key words: liver model, personalized, high precision, data-driven