Investigation of Detection Characteristics of Finger Pressure and Touch Area and Their Application to Pre-classifiers in Writer Verification

We propose a highly convenient authentication system, that requires a user to write a simple symbol on a touch-panel display of a smartphone and/or tablet terminal. The detection characteristics of the finger pressure and finger touch area on a touch-panel display are investigated since they are expected to be individual features that are independent from the written shapes. As a result, the exact pressure and contact area values are not detected but it is confirmed that the values corresponding to the pressure and the contact area are measured on the touch panel screen. Moreover, we propose to use the extracted features as pre-classifiers and apply them to writer verification. The verification performance is confirmed to be improved by the proposed classifier.