IEEE Trans Vis Comput Graph. 2014 Nov;20(11):1519-29. doi: 10.1109/TVCG.2014.2339835.
This paper presents a computational framework for modelling the biomechanics of human facial expressions. A detailed high-order (Cubic-Hermite) finite element model of the human head was constructed using anatomical data segmented from magnetic resonance images. The model includes a superficial soft-tissue continuum consisting of skin, the subcutaneous layer and the superficial Musculo-Aponeurotic system. Embedded within this continuum mesh, are 20 pairs of facial muscles which drive facial expressions. These muscles were treated as transversely-isotropic and their anatomical geometries and fibre orientations were accurately depicted. In order to capture the relative composition of muscles and fat, material heterogeneity was also introduced into the model. Complex contact interactions between the lips, eyelids, and between superficial soft tissue continuum and deep rigid skeletal bones were also computed. In addition, this paper investigates the impact of incorporating material heterogeneity and contact interactions, which are often neglected in similar studies. Four facial expressions were simulated using the developed model and the results were compared with surface data obtained from a 3D structured-light scanner. Predicted expressions showed good agreement with the experimental data.
本文提出了一种用于模拟人类面部表情生物力学的计算框架。使用从磁共振图像中分割出的解剖学数据，构建了一个详细的高阶（三次埃尔米特）有限元人体头部模型。该模型包括一个由皮肤、皮下层和浅表肌肉腱膜系统组成的浅层软组织连续体。在这个连续体网格中，嵌入了 20 对驱动面部表情的面部肌肉。这些肌肉被视为各向异性的，并准确地描绘了它们的解剖几何形状和纤维方向。为了捕捉肌肉和脂肪的相对组成，还将材料非均质性引入到模型中。嘴唇、眼睑之间以及浅层软组织连续体与深层刚性骨骼之间的复杂接触相互作用也进行了计算。此外，本文还研究了在类似研究中经常被忽略的纳入材料非均质性和接触相互作用的影响。使用开发的模型模拟了四种面部表情，并将结果与从 3D 结构光扫描仪获得的表面数据进行了比较。预测的表情与实验数据吻合较好。
