Extraction of feature vector based on wavelet coefficients for a palm print based biometric identification system

Palm prints are rich in features, can be easily acquired and hence different Palmprint features are used for the purpose of biometric identification. This paper presents analysis of wavelet coefficients for feature extraction of palm print images for biometry. A novel preprocessing technique for alignment of Palmprint images has been developed which allows the palm to be placed in any orientation in front of the camera. This paper also presents a method to extract a square shaped Region of Interest from the aligned image based on the palm geometry. Further processing has been done on the coefficients obtained from the Discrete Wavelet Transform to generate the feature vector. The method of feature extraction and the results obtained from feature matching using Neural Network have been discussed in this paper.