In this paper, we present a Cartesian-based co-ordinate system (CBM) to represent dorsal hand vein pattern and also introduce a novel method to extract the dorsal hand vein pattern that is based on the generalized method of moments. However, this method leads to high dimensional covariance matrix which is singular or rank-deficient. To overcome this problem, we use a fast and accurate algorithm to compute the inverse using the Moore-Penrose method. The biometric system developed is tested on a database of 100 images. The false acceptance rate (FAR), false rejection rate (FRR) and the matching time are being computed.
|