Design And Implementation Of Finger Vein Authentication System

Finger vein recognition technology is one of the world's most advanced biometric identification technology. Compared with other biometric technologies, finger vein recognition technology is highly accurate, highly security, high-availability, easy to use and so on, and gradually become a research hotspot in the field of feature recognition. In practice, however, the finger vein recognition technology is still urgently needed to overcome some of the difficulties. For example, in the acquisition process of finger vein image, due to the influence of external conditions such as light and fingers to their own circumstances and other factors, so that acquired images appear in fuzzy, uneven illumination, noise, displacement and rotation and other aspects of the problem, this will seriously affect the effectiveness of the finger vein recognition system.In order to improve the overall performance of the finger vein recognition system, we has established a three-dimensional model of the finger vein image acquisition, designed Edison hardware processing platform, presented a unique hierarchical matching algorithm, completed embedded operating system, designed application software, and finally completed the construction of the finger vein recognition system. Through joint software and hardware testing, we has proven scientific nature of acquisition terminal and processing platform, the uniqueness of the algorithm design and the practicability of the software system. Details are as follows:Firstly, we has established a 3D model of finger vein image acquisition. A unique optical design ensures that a finger vein image gray uniform and clear. A unique shell design can prevent the fingers from rotating and displacement. We has designed Edison hardware processing platform. Due to the parallel nature of SOC solutions, the system has a faster speed, higher integration and greater reliability.Secondly, we has presented a unique finger vein features hierarchical matching algorithm. The finger vein features are divided into both high and low levels. We has used the original coding algorithm and transformed feature extraction and matching process into the encoding process, so that the system has high real-time and high accuracy, and has stronger robustness for noise, rotation and displacement.Thirdly, we has completed the embedded operating system development. We have used C language to realize the equivalent MATLAB core algorithm and functions of image processing library. We has completed the application software development. It provides the background management and data storage mainly including registration, identification, inquiry, and settings.Finally, we has completed the construction of the finger vein recognition system. We have used 300000 pixels infrared camera, and got a clear vein image. Through the testing experiment for 64 groups of fingers vein samples, system identification speed reach 0.539 seconds, FAR and FRR respectively are 0.001% and 0%. The results show that the system can obtain high-quality finger vein image, and recognition algorithm has high recognition speed and recognition rate.