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：2015
卷号：3
期号：3
DOI：10.15680/ijircce.2015.0303086
出版社：S&S Publications
摘要：In recent years, multi-camera networks have become increasingly common for biometric andsurveillance systems. Multi view face recognition has become an active research area in recent years. In this paper,an approach for video-based face recognition in camera networks is proposed. Traditional approaches estimate thepose of the face explicitly. A robust feature for multi-view recognition that is insensitive to pose variations is proposedin this project. The proposed feature is developed using the spherical harmonic representation of the face, texturemapped onto a sphere. The texture map for the whole face constructed by back-projecting the image intensity valuesfrom each of the views onto the surface of the spherical model. A particle filter is used to track the 3D location of thehead using multi- view information. Videos provide an automatic and efficient way for feature extraction. Dataredundancy renders the recognition algorithm more robust. The similarity between feature sets from different videoscan be measured using the reproducing Kernel Hilbert space.
关键词：multi camera networks; face recognition; spherical harmonic; particle filter; Kernel Hilbert space
