Finger Vein Recognition Technology Research

Biometric recognition technology is a new discipline that accompanied by the development of computer technology, especially by image processing and pattern recognition technology. Finger vein recognition is a second-generation biometric recognition technology which is high-precision and high-speed. Since its inception, this technology has received great attention, and gradually into people's daily life, in every aspect of life has a wide range of applications.In this thesis, most of the research works focus on the image acquisition and finger vein image recognition algorithm. We compared the different lighting, different wavelength light sources, filters, and imaging devices and other equipment, choose the appropriate test device produced a prototype and collected finger vein images that in line with experiment requires. Finger vein recognition algorithm includes preprocessing, feature extraction, post-processing, registration and matching. In preprocessing we involves color images to grayscale images conversion and spatial smoothing. In feature extraction we test two algorithms. In post-processing we used binary morphological image processing methods, and improved classic bridge, refinement and deburring algorithm. In registration we use a registration algorithm based on grayscale histogram. In matching we use a matching algorithm based on Hausdorff distance. At last, we verify the images acquisition and vein recognition algorithm performance by a large sample of finger vein images library.