摘要
提出一种正面人脸关键特征点的自动提取方法。该方法选取两个瞳孔点、两个鼻孔点、鼻尖点、两个嘴角点共7个关键特征点进行自动标定的研究,通过在图像预处理阶段采用选择式掩模平滑算子和Fisher线性判别思想分别进行灰度图像的边缘检测和二值化处理,并结合灰度积分投影方法定位人脸区域,进一步应用人脸结构基本特征的先验知识定位面部主要器官,在此基础上结合Hough变换,SUSAN算子实现特征点的自动提取。实验证明:该方法能够准确、有效地标定人脸的关键特征点,具有一定的实用性,可应用于安全部门、身份证管理、电视会议等领域。
An approach is proposed to locate 7 key feature points on human front face automatically,including two pupil points, two nostril points,nasal tip,two corners of mouth points.By using selective masking operator and fisher＇s linear discriminant to distinguish the regions of facial color and non-facial color.Then fix on facial region,approximate positions of human features are determined by integral projection and knowledge of facial geometrical structure.At last,the positions of the feature points are extracted by methods of Hough transform and SUSAN operators,The experiments show that the method is reliable and practical.It can be applied to many different fields such as security departments,I.D.Card management,video conference,and so on.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第18期167-170,共4页
Computer Engineering and Applications
基金
国家自然科学基金No.60573179~~
关键词
人脸识别
人脸检测
灰度积分投影
二值化
face recognition
face detection
integral projection
binarization
