期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2013
卷号：2
期号：12
页码：3007-3010
出版社：Shri Pannalal Research Institute of Technolgy
摘要：The variation caused by illumination in faceimages is often greater than pose, facial expression andsometimes even between different individuals. In order tonormalize the effect of illumination for better face recognitionand to improve the face images to be more visually pleasing,three face image enhancement methods are devised. Theproposed methods are based on three stages namely thelogarithm transform, wavelet transform and wavelet basedimage denoising. Both qualitative and quantitative approachesare used to evaluate the performance of the three methods withface recognition using principle component analysis (PCA)being the main analysis in the quantitative approach. Theexperimental results obtained by testing images from theExtended Yale Face Database B shows that the wavelet basedmethods perform better than the conventional method.
