Research On Biometrics Arithmetic Based On Hand Geometry Features And Palmprint Features

1.IntroductionWith the rapid development of information technology, computer information to the community as a whole and networking, which makes information security and the importance of the unprecedented show. On various occasions such as: e-government, e-commerce, justice, access control, security, system control, anthropology, medicine, and so the need for accurate identification to ensure system security. Therefore, the people's identification technology requirements more stringent. Biometric technology is in these circumstances the right moment.Biometric technology is inherent in the use of the body for automated biometric identification technology, these features is the only life remain unchanged, reliability and stability. Now widely used in the biological biometric authentication technology not only fingerprints, palmprints, hand, face, iris, retina, hand and other physiological characteristics are voice, signature, and other behavioral characteristics. Against any single biometric for identity verification, when large database system, the recognition rate will decline. Well, how many people have a variety of characteristics, such as fingerprint, face, palmprints, iris, voice, retina, and other features, a comprehensive analysis is the identification of new areas of research. Manpower as biometric features, there are two characteristics of uniqueness and stability: Features hand and palmprints feature. In this paper the characteristics of the manpower is proposed based on simple hand and palmprints characteristics of the recognition algorithm.2. Major study contentsIn this paper, the relevant research results at home and abroad, in-depth understanding of the various Palmprint Identification System key and the corresponding method on the basis of the characteristics of fully integrating the staff in the pretreatment and identification method, etc. were taken relatively feasible approach, a useful exploration, and has been more optimistic about the rate of correct identification. The main research work include the following aspects:(1). Palm image preprocessing. According palmprint image contours in the region of the feature points physiological characteristics (convex and concave-shaped region), which is based on improved hand-shaped feature points palmprint effective regional positioning and segmentation method. Firstly, through scanning and thumb, in addition to the four fingers of the eight intersections to determine the bottom of the three regions. Secondly, the use of improved Valley Search method to find the three relatively stable (Valley) feature points, using characteristics to establish a coordinate system. Then coordinates with the horizontal distance from the perspective of image perspective correction. Finally once again the characteristics of the improved detection method to find hand feature points. Experiments show that, when not in the acquisition of images of palm registered a strictly limited circumstances, the paper proposed by the angle correction after the effective division of a more stable region.(2). Hand feature definition and extraction. After perspective correction, according to the three characteristics to establish a coordinate system, based on this definition of the five characteristics of hand, the thumb is in addition to the four fingers, the length and the width of the palm of your hand. Finger length is defined by the four fingers and means to get the root. Feature Extraction point fingers still used in this paper to improve the Valley Search, only to find upward, and that is characteristic root of the five fingers below the midpoint of feature points. Through the experiment, determine the width of the palm of your hand position.(3). Palmprint feature extraction. Using wavelet transform, to separate the palmprints of images multiresolution analysis. Improved wavelet energy characteristics of the composition, determined through experiments with greater ability to distinguish between right palmprints of 3,4,5-wavelet coefficients of wavelet energy features palmprints. Retain the image of palmprints feature information and reduce the amount of calculation at the same time, improve the eigenvector ability to distinguish between right palmprints.(4). Matching strategy and decision-making. A square difference distance were calculated hand and palmprints of feature matching feature matching, in the analysis of experimental data on the basis of the proposed matching of targeted normalized treatment, and response characteristics of the data collected using a dual-matching method , in the decision-making stage, through a decision-making experiment threshold T.(5). Experimental analysis. Palmprint set up their own library, based on this experiment designed palmprint identification, and the experimental results were deeper analysis. The experimental results show that this paper based on the characteristics of hand and palmprints feature fusion method has a high rate of correct identification, it is widely used.3. ConclusionPretreatment palm image recognition system is a key step to the back hand and palmprints feature extraction laid the foundation. Improved hand this feature points to find ways is simple, easy to achieve rapid characteristics, extracted a stable hand and palmprints effective regional characteristics. Palmprint images of texture images, and wavelet transform is the most attractive attributes of its multiscale edge or the edge of the multi-resolution capability described in this paper using wavelet transform palmprint extraction features, and improved through experiments palmprint wavelet energy feature the composition, and enhance the palmprint identification capability and reduce the amount of calculation. Matching stage in the decision-making, using the actual data for the treatment of the match and finalize the strategy. Experimental identified through the decision-making threshold, the results prove that, the paper by hand and palmprints double-feature matching method has higher accuracy rate, palmprint identification technology is a complement.