中文摘要：

多模态人脸识别技术是解决目前二维人脸识别领域瓶颈问题的一项重要手段。在二维人脸特征定位中结合Gabor变换和人脸弹性图方法，采用缩小搜索范围的方法提高系统效率；三维人脸特征定位首先采用PS方法进行粗略定位，然后仿照二维人脸识别方法定义Gabor变换系数向量进行精确定位，其中也采用一个模板来缩小搜索范围。试验证明，方法不仅效率高，而且能适应多种姿态和表情的变换。

英文摘要：

Multi-modal face recognition is an important way to solve the problem of 2D face recgnition. Gabor wavelet transform and Face Bunch Graph are integrated to locate the 2D face feature points, to improve the efficiency, a model of 11 × 11 is used to dwindle the range data. When to locate the 3D face feature points ,Point Signature is used to locate coarsely ,then Gabor is adopted to locate precisely. A model of 25 × 25 is also used to dwindle the range. Experimental results show that presented methods perform well and adaptive to pose and expression.