Orthognathic Soft-Tissue Prediction Based on Three-Dimensional Graphics Model Recovery

In this paper, we propose a novel algorithm for mandibular soft-tissue prediction to simulate the facial appearance post-operation. Specifically, the prediction problem is firstly cast to the framework of statistical deformable model recovery. To overcome the small sample size problem of traditional global statistical models, we present a Regional Orthognathic Statistical Deformable Model called Rosdem. Generated from the sample set of mandibular bone and mandibular surface, the prior-knowledge of basic morphology of the mandible and the soft-tissue is formularized of Rosdem and the variations of different individual samples are characterized as well. Then post-surgery appearance prediction is treated as a missing data problem, which is solved by a series of model recovery formulations. The prediction results is acceptable by the surgeries, and the experimental results show that the proposed algorithm achieves better performance than global model based methods.