中文摘要：

针对三维人脸数据庞大及识别效率低的问题，提出采用提取脊点及谷点表征人脸。脊点和谷点作为曲面局部区域内主曲率沿主方向变化的极值点，能够很好地表征三维人脸特征。对三维人脸提取脊点模型和谷点模型，通过对它们栅格化后生成对应的空间分布密度直方图实现人脸粗匹配，采用计算LTS－Hausdorff距离实现人脸的精确匹配。在GavabDB三维人脸库的实验结果表明，该方法具有较高的识别率。

英文摘要：

An approach for 3D face recognition is presented which uses ridge points and ravine points as features.Ridge points and ravine points,the extrcrna of the principal curvatures along their corresponding curvature lines,are powerful shape descriptors for shape matching.Ridge model and ravine model of 3D face are attained.The ridge model and ravine model are divided into grids, spatial distribution density histogram is created and used for face rough matching.LTS-HansdoriT distance is utilized for accurately matching the ridge points ＂model and the ravine point model of a given probe 3D face to those in the gallery. The results of the experiments on GavabDB show that the proposed approach have great capability for 3D face recognition.