Biometric Attendance System with Machine Learning Handwritten Signature Verification on Tablet

Handwritten Signature Verification (HSV) has been changed in the last decade as graphics tablets have come into widespread uses. A tablet can be used to capture a handwritten signature as samples of coordinate pairs. We propose a ML (Machine Learning) –HSV in a Biometric Attendance System (BAS). The first phase includes collection of 3 initial signatures for a certain student. We capture a set of 3 initial hand-written signatures from the user who can be identified according to the nine hand-written signature features: Number of Strokes, X coordinate sequence, Y coordinate sequence, Pressure sequence, Speed of Strokes, Angle of the first two strokes, Number of cross points, Proportion of Width and Length of signature, and Elapsed Time. In the second phase, we collect test signatures to do the authentication based on the 3 initial signatures for a certain user. The similarity is quantified. The higher value represents the better similarity. The application of BAS is to assist lecturer taking students’ class attendance record with their verified tablet signatures. The students’ attendance record will be evaluated against the students’ performance. The result will be educational to find out the impact of lecture to the students’ learning productivity.