An Identification System Based On Finger Vein Patterns

Traditional identification methods require passwords, keys or other external media. They are inconvenient and have the risk of being leaked or copied. Current identification technologies use human's biological characteristics to recognize people's identity, which greatly improves the convenience and security level of the identification process. Among many biometric-based identification technologies, finger vein identification technology has become a hot spot in current research because of its unique superiorities.In this study we completed the following work:(1) Introduced the background and significance of biometric identification technologies. Introduced several typical biometric identification technologies and summarized their advantages and disadvantages. Focus on the advantages and recent research developments in domestic and international of finger vein recognition technology.(2) Designed of a finger vein imaging system. Based on imaging principles and modalities, we designed the imaging system for acquiring finger veins and related mechanical structure. We designed the automatic light intensity adjustment system enables automatically handles lighting requirements under different light intensity needs.(3) Designed finger vein image processing algorithms. Finger vein images need to go through a series of image processing to extract features. In this study, based on the analysis of conventional finger vein segmentation algorithms, a multi-directional vein searching method based on the "valley-shaped distribution" of gray values is proposed.(4) Designed a template synthesis method controlled by a threshold, which can effectively eliminate the random variation information in the template, therefore to improve recognition accuracy.(5) Experimental verification. Experimental results show that the vein search method, template synthesis method and recognition method proposed in this study have good results. The identification system designed in this study has an accuracy of 99.10% in identification mode, an equal error rate of 1.03% in authentication mode, and the whole process takes less than 1 second. The performance of our system can satisfy the requirements of identification systems.