Research Of Hand-based Personal Automatic Identification Method Using Gabor Wavelet

Hand-based personal identification as a biometric technology possesses great potential of development, and it has great theoretical and practical values. Hand-based identification are gaining popularity and diverse applications for its characteristics of being non-invasive, low cost, and friendly use. In recent years, Hand-based identification technology has achieved unprecedented progress. But in the processes of its practical application its recognition precision cannot still satisfy our expected demands because the shifts and variations of hand or other disturbance exist when capturing the images. This paper conducts the systematic researches of the theory and technology of hand-based personal identification using two-dimensional Gabor wavelets for without any docking device system.In this paper, the commonly used hand-based identification theories and methods are briefly summarized. Research actualities, technology development and system design are also introduced in detail. According to the limitations of the existing location methods, we propose a method of key points based on one-dimension quadric spline"a trous"wavelet to provide eleven key points. We offer the method and result of feature areas location and preprocessing. Different from traditional palmprint identification system, this paper combines features of fingerprint and palm, adopts two Gabor wavelet filters groups on different frequency scales and orientations to extract features. Then we extract variance features based on Gabor wavelet features to form feature vectors of each feature area respectively. We apply the distance measurement based on inner product and matching fusion strategy to make the final decision. According to the system procedures, we conduct abundant experiments by using Matlab7.0 and provide the result of experiment. Comparing to the result of other existing methods, we prove the reliability and effectiveness of this system.In the end, we offer the outlook for the technique of hand-based personal identification. With the extensive applying prospect of hand-based identification technology, we also look forward to the increasing identification algorithms.