期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2019
卷号：7
期号：1
页码：21-26
DOI：10.15680/IJIRCCE.2019. 0701008
出版社：S&S Publications
摘要：For face recognition systems, impostors can obtain legal identity authentication by presenting the
printed images, the downloaded images or candid videos to the sensor. Since, the face is a unique biometric of the
individual and face recognition is the superior identification method. This paper proposes a novel method for face
spoofing detection by using features of Local Binary Pattern (LBP) and the popular machine learning approach SVM
used as classifier. The LBP is applied to each 3x3 matrix obtained from detected face through Viola-Jones algorithm to
get the features. The face image is segmented into number of different blocks and LBP Features are taken, then SVM
(Support Vector Machine) is used for determining whether the input image corresponds to live or fake face. Our
experimental analysis on a publically available NUAA face anti spoofing database following the standard protocols
showed good results.
关键词：Face;spoofing attack; Local Binary Pattern; NUAA dataset;
