Optimization approaches for soft–tissue prediction in craniofacial surgery simulation

Abstract

A system for interactive, 3–D, craniofacial surgery simulation is presented. It is used for the 3–D simulation of osteotomies of the facial and skull bones and for the prediction of soft–tissue changes caused by bone movement. The result of the simulation process is a 3–D, photorealistic model of the patient’s postoperative appearance that can be viewed from any position. The system is based on the individual preoperative bone structure of a patient’s skull derived from a computer tomography scan and on the patient’s photorealistic, preoperative appearance obtained by a laser scanner. The elasto–mechanical properties of the multi–layer soft–tissue are represented by springs. The model incorporates additional features such as skin turgor, gravity, and sliding bone contact. The prediction of soft–tissue deformation due to simulated bone movement is computed using an optimization approach. Several optimization methods have been tested and compared with regard to robustness of the simulation result and to computational costs. While the osteotomy simulation can be performed interactively, the computation of the corresponding soft–tissue changes usually takes less than 10 seconds even in sophisticated cases. Tests have been performed on a SGI O2 R10000, 175MHz. The system is able to simulate bimaxillary osteotomies, physiological jaw movement and has been used in the planning process in case of a craniosynostosis.