期刊名称：Signal & Image Processing : An International Journal (SIPIJ)
印刷版ISSN：2229-3922
电子版ISSN：0976-710X
出版年度：2011
卷号：2
期号：3
页码：69
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：In this paper, a multi-resolution feature extraction algorithm for face recognition is proposed based ontwo-dimensional discrete wavelet transform (2D-DWT), which efficiently exploits the local spatialvariations in a face image. For the purpose of feature extraction, instead of considering the entire faceimage, an entropy-based local band selection criterion is developed, which selects high-informativehorizontal segments from the face image. In order to capture the local spatial variations within these highinformativehorizontal bands precisely, the horizontal band is segmented into several small spatialmodules. Dominant wavelet coefficients corresponding to each local region residing inside those horizontalbands are selected as features. In the selection of the dominant coefficients, a threshold criterion isproposed, which not only drastically reduces the feature dimension but also provides high within-classcompactness and high between-class separability. A principal component analysis is performed to furtherreduce the dimensionality of the feature space. Extensive experimentation is carried out upon standard facedatabases and a very high degree of recognition accuracy is achieved by the proposed method incomparison to those obtained by some of the existing methods
