基于稀疏表示的快速三维人脸识别方法  

Method of Rapid 3D Face Recognition Based on Sparse Representation

作　　者：杨福猛[1] 袁晓斌[2] 相桂芳[3] 

机构地区：[1]三江学院电子信息工程学院,江苏南京210012 [2]安徽师范大学教育科学学院,安徽芜湖241000 [3]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《安徽师范大学学报（自然科学版）》2014年第6期543-548,共6页Journal of Anhui Normal University(Natural Science)

基　　金：国家自然科学基金(61371156)

摘　　要：随着计算机技术和三维成像技术的发展,三维人脸识别因不易受光照、装扮变化的影响成为人脸识别和身份验证的新趋势,但是对于表情、姿态变化其识别率还是有待于改善,时间开销较长.本文提出基于稀疏表示原理,对人脸重要的特征鼻尖点进行提取,采用最近邻分类器进行分类识别.实验结果表明,对表情变化等具有较高的鲁棒性和识别效果,且时间开销极小,优于传统的三维人脸识别方法.With the development of computer technology,3Dimaging techniques and 3Dface recognition becomes the new trend for face recognition and authentication because it is not easily affected by illumination and dressing change.But for expression and posture change its recognition rate still remains to be improved and a longer time is used.This paper,on the basis of the principle of sparse representation,extracts important features of nasal tip point and uses the nearest neighbor classifier to classify recognition.The experimental results show that it has high robustness and recognition effect of expression change and a little time is used.It is superior to the traditional method of 3Dface recognition.

关 键 词：三维人脸识别 稀疏表示 鼻尖点 最近邻分类器 表情变化 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]