Dorsal hand vein pattern feature extraction with wavelet transforms

Abstract

Abstract — Actually, hand vein biometrics is a recent technology that offers system for identification /authentication, it ranks among the best biometric modality by the results developed. Just like any recognition system this has four steps: the acquisition, enhancement, feature extraction and classification. This paper present the enhancement’s step of the SAB11 Data Base followed by new adaptive feature extraction method for the dorsal hand vein biometrics; which is the discrete wavelet transform. Keywords — dorsal hand vein patterns, wavelet transform, feature extraction . I. INTRODUCTION Hand veins Biometrics have received considerable attentions in recent years. With vein pattern offers one of the best results by their stability and unicity, still more, the biometrics of the hand veins are not expensive for realized and very convenient to use by users. Good recognition should have a good classification and a good classification should be above a perfect feature extraction phase this is where lies the strength of the biometric system, our work is focused on the dorsal hand veins feature extraction step, but the question asked is which method used to ensures a better feature extraction? In this paper the hand veins pattern are shown in gray level image. The main objective of this work is to provide a method which allows feature extraction of veins pattern from low quality images. II. PRIOR WORK They are several works about feature extraction of hand veins pattern, among them there is the Gabor filter, the Hough transform, discrete Curvelet transform, triangulation of minutiae...etc. most of his method are preceded by a preprocessing step where in the Gabor filter [1] and the Hough transform [2] they use the Median filter, Wiener in Gabor [1] and SIFT method [4], the Mexican hat in triangulation minutiae [5].The table 1 summarizes most of this works.