摘要
采用光学特征提取方法进行人脸识别中受到生物视觉特性的影响,导致人脸识别的精度不高,为了提高人脸识别的准确性,提出一种基于自适应模板匹配和三维人脸激光测距特征提取的人脸识别技术。首先在人脸候选区域建立人脸的几何结构模型,采用激光测距方法提取瞳孔与眼睑特征信息,用中值滤波器获得最佳的目标人脸图像灰度值,减少光照对人脸识别和特征检测的影响。然后对人脸进行边缘检测后提取三维人脸激光测距特征,用Adaboost算法建立人脸特征分类器,实现人脸识别。最后进行人脸识别的仿真实验,结果表明,采用该方法进行人脸识别,能提高模糊人脸特征的准确检测性能,提高人脸准确识别率。
Affected by biological visual feature in face recognition through optical characteristics extraction method,the face recognition accuracy is not high. In order to improve the accuracy of face recognition,proposes a face recognition technology based on adaptive template matching and 3 D face feature extraction based on laser ranging. First establish the geometric model of face in face candidate region. Extract the feature information of pupil and eyelid using laser ranging method. Obtain the optimal values of target face image by median filter. Reduce the illumination on face recognition and feature detection. Then extract 3D face laser ranging characteristics after edge detection. Realize face recognition through the establish of facial feature classifier based on Adaboost algorithm. Finally,the simulation results show that this method can improve the detection accuracy of fuzzy facial features thus improve the face recognition rate.
作者
张汉萍
ZHANG Hanping(Computer Department,Wuhan Polytechnic,Wuhan 430074,Chin)
出处
《激光杂志》
北大核心
2018年第7期87-91,共5页
Laser Journal
基金
湖北教育科学十二五规划课题(No.2015GB251)
关键词
人脸
激光测距
特征提取
识别
特征分类器
face
laser ranging
feature extraction
feature recognition
classifier
