Design And Implementation Of An Embedded Networked System Based On Finger Vein Recognition

Biometric personal identification techniques have attracted more and more scientist dueto its large potential security application. Compared with other biometric identificationtechniques, finger vein recognition is faster, more stable, and finger vein is moredifficult to be faked. In this paper, the author proposed some optimal methods for thefinger vein recognition algorithm on a DaVinci embedded board and realized anetworked system.This paper consists of four parts. Firstly, the author designed and fabricated threeprototypes of the collection devices for finger vein recognition. The author alsodeveloped a new strategy to design the adjustable infrared source, which canautomatically provide feedback of illumination intensities in order to generate desirableimages with distinct textures. Furthermore, the collection device can automaticallydetect whether the fingers of users are in the right place. Secondly, the author developedthe algorithm for image processing (image normalization, enhancement, denoising andsegmentation) and recognition (based on image matching and statistical learning).Thirdly, the author proposed packages and Codec Server which are compatible for thexDAIS/xDM for the finger vein recognition algorithm on the TMS320DM6446board.The multi-thread application program is used and has achieved the functions of userinterface, and control, capture, display and database management. Lastly, the authordesigned a networked system based on finger vein recognition. The system consists ofepigenous machine and several embedded boards. They communicate through LAN.The contributions of this paper are as follows: Designed several collection devicesand an algorithm for finger vein recognition. The image is clear and the algorithm hashigh accurate rate. Transplanted the whole algorithm for finger vein recognition intoembedded systems. Designed a networked system based on finger vein recognition.