ABSTRACT
In this paper, we propose a hand vein recognition algorithm based on reference points and the NMI (normalised moment of inertia) of vein images. Firstly, those reference points with highest matching credibility are selected according to the relative distance between the feature points, including the intersection points and endpoints, after vein image preprocessing. Then the angles between the adjacent connections of the reference point and feature points are calculated in order to recognise vein images in the first place. Secondly, the general NMI values and sub-block NMI values of vein images are extracted as characteristics to enable further recognition. With both global features and local features of vein images, the information about vein characteristics could be fully utilised. There is little redundant information in the reference point algorithm, so the speed of recognition is fast and real-time. Using the NMI of vein images as a characteristic overcomes the
1 INTRODUCTION
With the development of science and technology, effective protection of information safety becomes increasingly important. The technology that could recognize an individual’s real identity effectively is urgently needed. Using the body’s innate biological characteristics to recognize identities safely and consistently, biological recognition technology has developed significantly in recent years. Compared with other biological characteristics, such as fingerprints, palm prints, irises, facial images, voice, gait, etc. (Menotti et al. 2015; Abo-Zahhad et al. 2014), hand vein recognition technology has many advantages, such as being non-contact, uncopiable, having stable characteristics, and requiring a simple sampling device (Hu et al. 2014).
