Research Ofdual-modalbiometric Recognition Based On Fingerprint And Finger-vein On The Match-score Level

With important information in biometric traits, Biometric Identification has become an important way for individual identification. By the development of this technology, single biometric has some problems in the aspects of universality and accuracy, which can be solved by multi-modal biometric identification system. So, the technology of dual-modal recognition fusing finger print and finger-vein in the match-score level are researched in this paper.Since the development of fingerprint is very mature, this paper mainly introduces finger vein, including pre-proceeding, feature point extraction and matching, implementation of dual-mode biometric system. After getting finger vein image, ROI location, Normalization, image enhancement and segmentation can be done with different algorithms, which are compared and analyzed by Matlab simulation.The feature extraction of finger vein is based on minutiae, which is different from the finger print, and the research shows that among the minutiae, bifurcation points carry the most important information. To get this information, a new algorithm is used to get the minutiae's location and angle, this algorithm can make up for image rotation and offset and improves the identification result.In the match-score level fusion module, different fusion level, different score normalization model and matching algorithms are introduced and compared. According to the result, Weighted Average algorithm is used in this paper and GEC is used to improve this algorithm, it can optimize the weight assigned to different biometric modalities, based on which the match-score fusion of finger print and finger vein is implemented. the simulation show that, the dual-modal biometric system in this paper not only increase the result and security level, but also the weight assigned by it can make a corresponding change in different environment with strong adaptive ability.