Özden Niyaz, Zehra Gülru Çam, and Tülay Yıldırım
Identification, pattern recognition, image processing, image recognition
With rapid advances in technology, biometric control systems have become part of the contemporary world. Wrist pattern is a biometric that is unique to the individual and can be seen on the skin of most individuals. This study focuses on the left wrist vein images of 20 people, in a visible light band, using a simple smartphone 5 MP camera. Selecting a reference line on wrist wrinkle, the wrist vein regions are cropped as square. The red layer of the RGB vein image has been filtered with Gaussian Low Pass Filter to eliminate any high frequency noise. Sharpening and contrast stretching operations have been used on the filtered images. Sub-pixel correspondence matching, using phase-only correlation method, can yield efficient results for biometric image matching. In this paper, phase-based correspondence matching pattern recognition method is used in order to perform person recognition from wrist vein patterns, which have been captured by a smartphone camera. It is argued that, the experiments this study has conducted demonstrates that phase-based correspondence matching provides an effective means for wrist vein recognition.