Handwriting Analysis Goes 3D
Forgers beware. A new technique that uses three-dimensional holograms to analyze handwriting samples exposes writing characteristics that forgers can't fake. The method may prove to be the most powerful tool yet in identifying fraudulent signatures on checks and other legal documents.
Counterfeiters cost businesses and consumers billions of dollars a year by faking signatures on everything from wills to credit card receipts. Traditionally, forensic handwriting experts have tried to spot forgeries by analyzing the sequence of pen strokes used by the author to create a word. But experts often have a hard time discerning these "stroke dynamics," especially if a skilled forger is at work.
Scientists at the Università degli Studi “Roma Tre” in Rome tried to improve on existing techniques by using a hologram generator that creates 3D images of writing samples. As reported in the 10 August issue of Journal of Optics A, the device transforms seemingly flat letters into landscapes of hills and valleys that reveal the pressure and stroke sequence used to create each word. For example, when strokes made with a ballpoint pen cross each other, the second stroke forms a bump over the first. "This tells us the order of stroke crossing--something you wouldn't be able to see on a microscope," says Lorenzo Cozzella, an electrical engineer and co-author of the study. Because forgers only mimic the look of a signature, 3D analysis allows forensic experts to distinguish between fact and facsimile. The researchers tested their system by comparing writing samples made with various combinations of pen and paper types. They found that the holographic image indicated the proper stroke order in almost 90% of cases.
This is the first time holograms have been used to analyze handwriting, says Zeno Geradts, a forensic scientist at the Netherlands Forensic Institute in Rijswijk. But he believes the researchers should directly compare their holographic method with current handwriting-analysis techniques to see how the new approach stacks up. "If it really works, it would be a valuable tool for forensics," says Venu Govindaraju, a pattern recognition expert at the University at Buffalo in New York. "This could go a long way in helping us spot forgeries."