摘要
研究人脸识别问题,当人脸的表情发生很大变化时,提取的三维几何特征会发生较大变化,造成识别准确度很低,影响识别效果。为解决上述问题,提出了一种改进几何特征的三维人脸识别方法。通过求出人脸几何特征差异的关联程度,计算哪种三维几何特征变化与识别结果是无关的,哪种与识别结果是关联的,解决了人脸表情变化时的准确三维识别问题进行仿真。结果表明,改进方法实现识别高效,可使得三维人脸识别的准确度大大的提高,为设计提供了支持。
This paper studies the problem of 3d face recognition.The recognition accuracy of extraction of 3d geometric features,when the face expression changes greatly,is very low.This paper puts forward a 3d face recognition method based on improved geometric features.Through calculating the connection degree of face geometry characteristic differences,if the 3d geometric characteristics changes are extraneous to the recognition results or not is calculated,solving the problem of accurately identifying the face expression changes of three-dimensional.Experimental results show that,the method is simple,efficient,and enhances the 3d face recognition accuracy greatly.
出处
《计算机仿真》
CSCD
北大核心
2011年第4期291-294,共4页
Computer Simulation
基金
河南省教育厅自然科学研究计划项目(2010A520025
2009B520014)
河南省重点科技攻关计划项目(102102210409
102102210176)
关键词
三维人脸识别
几何特征
差异关联程度
Three-dimensional face recognition
Geometric features
Connection degree of differences
