Feature extraction based DCT on dynamic signature verification

Signature verification is one of the most accepted biometric techniques, because a signature is a part of everyday life, although less accurate than biometric techniques such as using the iris. In this field, much attention has been paid to features, because a verification system should be able to overcome problems such as forgeries, insensitivity to intra-personal variability and sensitivity to inter-personal variability. In this paper, we present a simple and efficient approach to on-line signature verification, based on a discrete cosine transform, which has been applied to 44 time signals, such as position, velocity, pressure and angle of pen. Experiments are carried out on two benchmark databases, SVC2004 and SUSIG. The forward feature selection algorithm is used to search for the best performing feature subsets. The proposed system is tested with different classifiers, with skilled forgery, and equal error rates were 3.61%, 2.04% and 1.49% for SVC2004 Task1&2, Task2 and SUSIG databases, respectively.