Hand geometry verification using time series representation

Biometric authentication based on human physical traits has recently been heavily studied; these biometric sources include face, hand geometry, voice, fingerprint, iris, retina, etc. The hand geometry is one of the most conventional biometric since it is fairly easy to implement and acquire the data, comparing to other biometrics such as retina, iris, or DNA sequences. In this work, we propose a novel time series representation for hand geometry system by converting raw images into time series data, where this representation can gracefully handle variability of hand's position, translation, and rotation, especially in a peg-free system with the help of a Dynamic Time Warping similarity measure. We demonstrate the utility of our approach by implementting the real hand geometry verification/identification system, and it has proven to work effectively and competitively with low false acceptance and false rejection rates.