Advances in collision detection and non-linear finite mixed element modelling for improved soft tissue simulation in craniomaxillofacial surgical planning.

BACKGROUND

There is a huge demand to develop a method for assisting surgeons in automatically predicting soft tissue deformation in terms of a bone-remodelling plan.

METHODS

This paper introduces several novel elements into a system for the simulation of postoperative facial appearances with respect to prespecified bone-remodelling plans. First, a new algorithm for efficient detection of collisions, using the signed distance field, is described. Next, the penalty method is applied to determine the contact load of bone on facial soft tissue. Finally, a non-linear finite mixed element model is developed to estimate the tissue deformation induced by the prescribed bone remodelling plan.

RESULTS

The performance of the proposed collision detection algorithm has been improved in memory requirements and computational efficiency compared with conventional methods. In addition, the methodology is evaluated over both synthetic and real data, with simulation performance averaging <0.5 mm pointwise error over the facial surface in six mid-face distraction osteotogenesis procedures.

CONCLUSIONS

The experimental results support the novel methodological advancements in collision detection and biomechanical modelling proposed in this work.