期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2015
卷号：4
期号：12
页码：12498
DOI：10.15680/IJIRSET.2015.0412154
出版社：S&S Publications
摘要：Facial recognition is an approach or a system to identify any individual. It can done by analysing andpattern comparison. It is mainly used to provide the security for an individual. It faces some difficulty in case ofdifferent illumination. To overcome this, invariant illusion algorithm is used in this paper. Due to this,our system canwork well under a different circumstances. It followed by three stages; preprocessing, a feature extraction, andrecognition. Wavelet-based normalisation technique lightens the dark and highlight areas of the face image. To imitatethe human visual system and examine a face image from the internal facial features to the external facial shapes,multiple face models along with complementary features are used. Features are extracted using Fourier transform. Toclassify the image, a simple metric Euclidean Distance is used.
关键词：Euclidean distance; Fourier feature; PCLDA; Single Scale Retinex; Wavelet-based normalisation
