Text-dependent User Verification of Handwritten Words and Signatures on Mobile Devices

摘要

In this work we propose an online verification system for both signature and isolated cursive words that uses histograms of angles, velocity and Fourier descriptors. The proposed system is designed to be used in a mobile device with limited computational capability and where users often draw using their fingertip instead of a stylus pen. The goal is to use signatures or isolated words to unlock certain applications, content or the device itself. In this context, isolated words have certain desirable properties that can be more useful than signatures. Different isolated words can be used to verify the user in different applications, combining knowledge-based security systems (i.e. passwords) with a behavioral biometric verification system. The proposed system creates a profile of several histograms sets, Fourier descriptors and statistical information based on a number of input samples. A simple Manhattan distance measure is used to compare the query sample with the profile in order to verify the identity of the user. This method can achieve an equal error rate as low as 3.25% for isolated words.