摘要
随着火车票实名制的不断推广,人工核实身份的验票方式已不能满足实际需求,鉴于此本文提出一种基于词包模型的人脸身份认证算法,通过人脸比对自动完成身份核实.首先提取每幅图像的尺度不变特征变换(SIFT)描述子;其次利用词包模型(BOVW)构建人脸的典型特征;随后训练SVM分类器,将同一人不同年龄段的图像作为同一类,针对同一人的类内相似性和不同人的类间差异性进行建模;最后通过SVM分类器分别对旅客图像和其身份证图像进行分类,根据所属类别的一致性判断是否属于同一人.实验结果表明,本算法能有效地进行身份认证,并且针对图像质量较低、光照情况不可控的情况仍可达到比较高的准确率.
With the development of train ticket real-name system,manual identity verification method no longer meets the actual demand.Therefore,this paper proposes a face verification method based on the bag-of-visual words model.This approach aims to automatically identity the verification based on face matching.First,the scale invariant feature transform（SIFT） feature is extracted for each image.Then,the bag-of-visual words model is used to construct the face typical feature.The study also trains the SVM classifier,which models the similarity between two images of an individual and the difference between two individuals.Finally,the passenger image and his ID card image are classified to determine whether they are belonging to the same person or not.The experimental result indicates that the proposed algorithm is able to effectively accomplish face verification,especially perform well in the circumstance with low image quality,and uncontrollable light and bright.
出处
《交通运输系统工程与信息》
EI
CSCD
北大核心
2013年第3期85-90,共6页
Journal of Transportation Systems Engineering and Information Technology
基金
铁道部专项资金支持课题研究项目(1119DZ4303)
铁道科学研究院基金资助项目(1151GC1103)
关键词
智能交通
人脸认证
词包模型
火车票实名制
SIFT特征
intelligent transportation
face verification
bag-of-visual words
real-name system for ticket
scale invariant feature transform（SIFT） descriptors
