期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2009
卷号：9
期号：12
页码：223-230
出版社：International Journal of Computer Science and Network Security
摘要：The present paper proposes a novel scheme of face recognition system based on texture features derived from co-occurrence parameters using 1-level linear wavelet decomposition technique. The present method divides the face into four parts. The novelty of the present scheme lies on usage of distance function on texture features for the recognition of facial images. The distance function evaluates the distance between each individual texture feature and eliminates the retrieved facial images in a sequential manner. By this the proposed paper expands the scope of structural face recognition system towards reliable and computationally inexpensive model. The proposed method is implemented on a large facial database collected from FGnet aging database and Google Images. The recognition rates between various linear wavelet transforms and with the original image are compared.
关键词：Pose; Illumination; Feature extraction; Orientation; class seperability
Loading...
