Signature is one of the oldest and most common form of biometric identification in these days, thanks to that it’s fast and easy to make, and the most important: unique. That’s why the signature is organic part of our days, from contracts to any statement or bank transfer we have to place our sign on these documents to consider valid.
However as in most areas of life, we must be count with swindlers, who try to copy these signatures and commit forgery with this to achieve their goals, and harm others. Therefore, we need supervision, that has been performed by forensic experts manually for years.
In our days it’s natural, that machines and software’s exist to perform most of the tasks, with which the companies can save time and money. That’s true in signature verification’s world. Today in many places the signature recording is electronic and can be therefore efficiently checked by automated systems. Unfortunately, these system’s accuracy and reliability is far from forensic experts, but solution are continuously evolving.
My thesis goal is to implement such a signature verification system, that can perform verification based on given samples. So it’s determine that the signature is from the owner with the best accuracy.
I made my program in Java programming language with the help of Eclipse development environment. I use the ImageJ tools to process the signatures, to which I will apply the presented algorithms and reach the desired result. Finally, I compare these results with other trustworthy sources to get a real picture of my system’s performance.