基于小波变换和SVM的眉毛身份验证系统研究  

Eyebrows verification based on wavelet transform and SVM

作　　者：曹俊彬[1] 李玉鑑[1] 黄琰[1] 

机构地区：[1]北京工业大学计算机学院,北京100022

出　　处：《计算机应用研究》2009年第11期4202-4204,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60775010);北京市属市管高等学校中青年骨干教师培养计划资助项目PHR(IHLB)

摘　　要：为了对眉毛这一新颖生物特征开展识别研究,提出了一种基于小波变换方法和支持向量机(SVM)的眉毛身份验证方法。其基本思想是用小波变换提取眉毛图像特征,然后用SVM进行训练和验证。在自建的100人眉毛数据库中进行的实验结果表明,该系统具有较低的错误拒绝率29.58%和错误接受率8.22%,从而验证了眉毛用于个人身份鉴别的可能性和有效性。In order to study the novel biometric of eyebrow, this paper presented an eyebrows verification based on wavelet transform and support vector machines （SVM）. Extracted the features of the eyebrows image by wavelet transform, and then classified them based on SVM. Verification results of the experiment on an eyebrow database taken from 100 of self-built personal demonstrated the effectiveness of the system. The system had a lower FAR 8.22% and FRR 29.58%. Therefore, eyebrow recongnition may possibly apply to personal identification.

关 键 词：生物特征识别 眉毛识别 身份验证 小波变换 支持向量机 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]