Research Of Palmprint Feature Extraction Technology

With the fast development of computer technology in recent years, biometrics technology as a new identification technology, has been widely used in government, finance, military and other aspects. The traditional identification methods have been unable to meet the information security requirements, biometric technology comes into being under such conditions. Biometrics refers to the use of the inherent characteristics of the organisms (iris, face, fingerprint, palmprint) or behavioral traits(signature, gait, gestures, etc.) a technique for authentication. Palmprint recognition as an emerging biometric technology, has the incomparable advantage of other biometric technologies, including acquisition equipment, stability, uniqueness, easy to accept etc. With theses advantages, palmprint recognition by the more widely used.Palmprint identification system consists of three main parts:image acquisition and preprocessing, feature extraction and feature matching. In this paper, the feature extraction algorithm, one of the key technologies on palmprint identification system is researched. The main content is as follows:1. Through access to relevant references, research of feature extraction in detail.2. Canny operator to extract the main line of palmprint images and using morphological dilation processing on the basis of an improved thinning algorithm. The proposed algorithm can solve the problem such as excessive corrosion fracture, spikes and non-single-pixel when thinning the main line.3. Research of the existing feature extraction algorithm, an palmprint texture feature extraction algorithm based on non-sampling Contourlet transform and two-dimensional principal component analysis is proposed. In this algorithm, the processed palmprint image is processed by non-sampling Contourlet transform, decomposition to get the coefficient matrix of the image,and then use two-dimensional principal component analysis to extract the palmprint image feature vector. Finally the minimum distance classifier is used to classify and identification. The experimental results show that the method has better features performance and higher recognition rate.