摘要
移动人脸识别过程中受到人脸反射亮点的漂移影响,导致定位识别的准确度不高,为了提高移动人脸识别的精度,提出一种基于激光光点定位技术在移动人脸识别方法。采用激光光点定位技术进行移动人脸的面部图像特征采样,对采集的人脸图像进行面部姿态矫正和模板匹配处理,采用图像二值化处理方法降低光点对移动人脸识别的干扰影响,在二值化图像中采用加权平均方法进行人脸边缘轮廓检测和面部区域化分割,提取反映移动人脸差异化信息的特征量,采用激光光点定位进行人脸特征定位识别,在激光光点高亮区域提取出脸部特征的全部关键点,实现人脸识别。仿真结果表明,采用该方法进行移动人脸识别的准确识别概率较高,抗干扰性较强,人脸特征点平均定位时间较短,实现高效快速的人脸识别。
The drift of the face reflection bright spot during moving face recognition,resulting in recognition accuracy is not high,in order to improve mobile face recognition accuracy,proposed a laser spot positioning technology based on mobile face recognition. The laser spot positioning technology is used to sample the facial image features of the moving face,and the facial image correction and template matching processing are performed on the collected facial image,and the image binarization processing method is used to reduce the interference effect of the light spot on the moving face recognition. In the binarized image,the weighted averaging method is used to perform face edge contour detection and face region segmentation,and the feature quantity reflecting the moving face differentiation information is extracted. The laser spot location is used to perform face feature location recognition in the laser spot. The highlight area extracts all the key points of the facial features and realizes face recognition. The simulation results show that the method has higher accurate recognition probability,stronger anti-interference ability,shorter average location time of face feature points,and achieves efficient and fast face recognition.
作者
杨鑫
王爱学
YANG Xin;WANG Aixue(Wuhan Polytechnic,Wuhan 430074,China;School of Geodesy and Geomatics Wuhan University,Wuhan 430079,China)
出处
《激光杂志》
北大核心
2018年第12期116-120,共5页
Laser Journal
基金
湖北省教育厅科学研究计划项目(No.B2013267)
关键词
激光光点定位
移动人脸识别
二值化图像
识别率
laser spot location
moving face recognition
two valued image
recognition rate