摘要
提出了一种基于特征加权聚类的视频人脸图像识别方法。该方法指定样本人脸图像特征点,通过增加特征点多方向窗口加权计算,使得中心特征点四周各点具有相异权重,利用加权聚类人脸识别方法检测限制搜索区域,最后根据每一个侯选区域的几何信息及人脸特征验证是否为指定人脸。实验结果表明,提出的算法实现简单、误检率低、检测速度快,适合实时视频监控取证系统应用。
A human face recognition method based on feature weighted cluster for video forensics is presented.This method assigns the sample person face image characteristic point,through computing multi-direction window weighting of the increase characteristic point,enables central characteristic point all around each spot to have the different weight,searches the region using the weighting cluster person face recognition methods examination limit,finally acts according to each candidate region the geometry information and the human face characteristic confirms whether the sample human face.The experimental results show that the proposed algorithm has high speed and low error-detection rate,so it can be used in the real-time video surveillance forensics system.
出处
《计算机与数字工程》
2010年第10期125-128,154,共5页
Computer & Digital Engineering
基金
湖南省教育厅项目(编号:08C027)
2007湖南省公安厅科研计划项目批准立项课题资助
关键词
视频取证
特征抽取
加权聚类
人脸识别
video forensics
feature extraction
weighted cluster
face recognition
