A simple and effective method for online signature verification

Abstract

This paper presents a simple and efficient method for online signature verification. The technique is based on a feature set comprising of several histograms that can be computed efficiently given a raw data sequence of an online signature. The features which are represented by a fixed-length vector can not be used to reconstruct the original signature, thereby providing privacy to the user's biometric trait in case the stored template is compromised. To test the verification performance of the proposed technique, several experiments were conducted on the well known MCYT-100 and SUSIG datasets including both skilled forgeries and random forgeries. Experimental results demonstrate that the performance of the proposed technique is comparable to state-of-art algorithms despite its simplicity and efficiency.