利用对称性实现三维人脸姿态的自动估计*

描述了一个能够快速精确地对三维人脸姿态进行自动估计的系统，提出了利用人脸的反射对称特性自动估计三维人脸姿态的方法，通过扩展高斯图像及最小包围球来得到三维人脸对称平面，利用搜索得到的鼻尖顶点对人脸进行估计，然后对估计在规定范围内进行修正，最终得到精确的估计结果。以三维扫描仪扫描的真实人脸数据作为输入对系统进行了验证，实验表明该方法不但具有很好的精确性和鲁棒性，而且能够很好地应用到实际应用中。

Estimating 3D facial pose with symmetry

This paper described a system that could be used to estimate 3D facial pose rapidly and accurately, and proposed a robust pose estimation method for 3D face models,which was based on facial reflective symmetry. With extended Gaussian image and smallest enclosing ball theories, the method detected the symmetry plane of 3D face at first, and searched the tip of nose, then estimated the pose approximately. In order to obtain more accurate result, the method corrected the approximate estimation in a given confine. The experiments using real face data obtained by 3D scanner, demonstrated the effectiveness as well as robustness.