Dorsal hand veins based person identification

Biometrics is a way that identifies people with the help of physical human features. There are many ways of bio-metric identification and recognition systems such as fingerprints, face, iris and veins etc. However, these conventional methods have some problems with respect of performance and convenience. Every human hand has unique veins patterns. Hand veins based recognition is most feasible than all of other conventional methods especially because of its easy acquisition process and also difficult to forge hand vein pattern. Patterns are taken from inside the body rather than obtaining from outside the body. Due to no physical contact, internal features and patterns from live body makes it more secure than other methods. In this research, we present a new method for person identification based on hand veins. The proposed system consists of pre-processing, vein enhancement and segmentation, feature extraction and finally matching. A new filter bank based method for hand veins enhancement is presented here. The proposed system is tested and evaluated using Bosphorus hand vein dataset which consists of 1200 hand images from 100 different people with 12 images per person. The proposed system has achieved 1.3% false acceptance and 1.75% false rejection rate respectively at a threshold of 0.85. Overall accuracy achieved by proposed system is 96.97%.