脸部特征点定位方法综述  

Review to facial features localization approaches

作　　者：韩玉峰[1,2] 施铜兴[1] 王小林[1] 

机构地区：[1]安徽工业大学计算机学院,安徽马鞍山243002 [2]安徽工业大学数理学院,安徽马鞍山243002

出　　处：《计算机工程与应用》2012年第1期180-182,218,共4页Computer Engineering and Applications

基　　金：2011年安徽高校省级自然科学研究项目重点项目(No.KJ2011A040)

摘　　要：人脸识别技术极大推动了图像处理、模式识别、计算机视觉等诸多学科的发展。人脸部特征点的定位是人脸识别中的关键步骤,定位准确与否直接关系到后续应用的可靠性。系统综述了特征点定位六大类方法,分为基于灰度信息、先验规则、几何形状、统计模型、小波和3D方法,并给出了对各方法的性能评价以及对未来的展望。Face recognition technology has greatly promoted the development of image processing, pattern recognition and computer vision. Human facial features positioning is a key stage in face recognition and the accuracy of the positioning directly relates to the reliability of subsequent applications. This paper systematically reviews the six categories of methods in facial features positioning, namely, those based on grey-level information, a priori knowledge, geometric shapes, statistic models, wavelet and three dimensions. It evaluates the above-mentioned methods and expresses prospects.

关 键 词：人脸识别 人脸特征点定位 脸部特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]