MonogenicCode: A Novel Fast Feature Coding Algorithm with Applications to Finger-Knuckle-Print Recognition

Biometrics based personal authentication is an effective method for recognizing a person's identity. Recently, it is found that the finger-knuckle-print (FKP), which refers to the inherent skin patterns of the outer surface around the phalangeal joint of one's finger, can serve as a distinctive biometric identifier. In this paper, a novel feature extraction and coding method, namely MonogenicCode, is presented based on the monogenic signal theory, and is applied to FKP recognition. For each image pixel, the associated MonogenicCode is a 3-bits vector obtained by binarizing the monogenic signal at this position, and it can reflect the local phase and orientation information at that position. Experiments conducted on our established FKP database indicate that this new method achieves competitive verification accuracy with state-of-the-art methods, while it needs the least time for feature extraction, making it the best choice for real-time applications.