It is my belief that the peculiarities and distinctions in each individuals’ handwriting can be uniquely identified and classified using new techniques made possible by Tablet PCs. Tablet PCs have the hardware to capture a multitude of handwriting data that go far beyond just the path of a stroke. Such hardware is able to decipher information about how the user is gripping the pen even before it touches the screen and afterwards, it is able to collect data about how fast someone writes, how much pressure they apply when writing, and even how someone holds their pen.  Unlike classical forensic handwriting techniques that analyzed only the visual strokes of a sample, we now have access to an entirely new dimension in the sample to use in identification.

My research seeks to capitalize on the Tablet PC’s hardware capabilities of collecting much more information about an individual’s handwriting and coupled with machine learning techniques and algorithms, it is my goal to be able to create highly accurate handwriting identification models for a variety of multiuser or biometric applications.
Accurate handwriting identification models have a wide range of possible applications especially now and in the future as new devices with touch screens and pen inputs become more prevalent.
  The most obvious application is in security and biometrics as we already use signatures as a form of verification and identification, although not very securely. Many grocery and retail stores already collect signatures electronically using pen inputs, but if these were linked with handwriting identification models, it would provide real time feedback telling the store if the person signing is really who they are signing as. And even everyday tasks such as logging into a computer can be made much more secure with handwriting identification since a password can be stolen and easily used by anyone where it would take great practice and skill to match someone’s handwriting as well.
Multi-user workstations
  Further applications can be for multi-user PCs where the PC will recognize your text input and adjust settings and preferences automatically for you.
  A few classrooms in the US have began outfitting classrooms with laptops for each student, if handwriting identification was applied in such a setting, it could catch and deter cheating as it will be able to identify work to the actual student that wrote it no matter under whose identity it is submitted as.

Zuye Zheng | Ananda Gunawardena