Palmprint Authentication Using Time Series

摘要

Automated personal authentication using biometric features is getting more and more popular for solving the security problems. A new branch of biometric technology, palmprint authentication, has attracted increasing amount of attention because palmprints are abundant of line features and thus low resolution images can be used. In this paper, we propose a new approach for palm-print feature extraction, template representation and matching. Using of time series technologies such as SAX representation and MINDIST calculation is the key to make this new approach simple, flexible and reliable. Experiment shows that this approach can achieve an accuracy of 98.7% when performing one to one verification on a 600 palmprints database. This new approach, which is very computationally efficient, also facilitates the biometric feature fusion as well as palmprint identification using incomplete templates.