摘 要： 提出了一种基于改进 Ha u s d o r f f距离的人脸相似度匹配的方法， 该方法首先将人脸划分为脸型、 双眼、 鼻、 嘴等几个特征点

集， 分别计算各部分的改进 Ha u s d o r f f 距离， 然后进行加权计算相似度。利用该方法， 在 A S M( 主动形状模型) 定位人脸的基础上进

行了人脸检索。 实验表明， 利用人脸相似度计算方法对人脸特征库进行搜索， 达到 了较好的效果。同时结合 A S M 自动人脸检测， 本

方法可以全自动完成人脸匹配， 应用于人脸识别及数字娱乐等领域。-Abstract: This paper presents a distance based on an improved Ha usdorff face similarity matching method, which will first face is divided into face, eyes, nose, mouth, and several other feature point sets, were calculated to improve the various parts of Ha usdorff distance, and then weighted similarity. Using this method, in the ASM (active shape model) positioned on the basis of human face human face retrieval. Experiments show that the use of face similarity calculation method for facial feature database to search, to achieve good results. ASM automatically combined with face detection, this method can be completed fully automatic human face matching, used in face recognition and digital entertainment and other fields.