|
Biometric Standards, Performance, and Assurance LaboratoryDepartment of Industrial Technology |
 |
 |
Dynamic Signature Verification This research is being led by Dr. Elliott
Dynamic signature verification is a subset of that larger science that includes fingerprint recognition, hand geometry, and voice recognition. Signature verification is primarily behavioral in nature like voice recognition, but has some very unique traits which make it harder to test and evaluate. These challenges include the fact that a signature is learnt, it contains variant measures, it can be changed by the owner of the signature, and that a signer might have several versions of the signature, depending on the intent of the signer. Current research includes providing an assessment of Dynamic Signature Variable Traits in Signature Forgery; Perception of Signature Strength, Repeatability of Signatures (see figure to the right), and an assessment of force on a digitizer. Other work includes testing protocol design for dynamic signature verification. Dynamic Signature Variable Traits in Signature Forgery centers on the fact that the signature may not be verified at that specific moment (unlike the other biometrics), but may be validated at a later date. Furthermore, understanding an impostor distribution is also a challenge in the fact that other biometrics use a zero-effort attempt, “where an impostor uses his or her own biometric sample and claims the identity of a different enrollee” (WG1, 2005). Thus, dynamic signature verification is unique among other biometric authentication methodologies as there is no clear defined way of creating a forgery. This research examines two aspects of a forgery – the first is the perception of the signature to forgery (how easy an individual perceives the signature to be forged), and the second is the amount of knowledge that a forger has about a signature. The dynamic variables of the signature were then examined to establish which statistical variables were susceptible to forgery using forensic tools. For dynamic signature verification, a zero-effort attempt would cause the forger to write their own name instead of that of the target. The Perception of Signature Strength is another important aspect in understanding the vulnerability of a signature. For example, a forger has access to a number of signatures, some of which they perceive to be difficult as opposed to an "easy" signature. This research attempts to define the strength of a signature, and then to analyze that signature dynamically to estimate the variables that are subjected to weekness. Signatures were redistributed for individuals to rank according to difficulty of forgery and to include reasons on why certain signatures would be difficult to forge. The data was coded to allow for analysis of patterns which indicate what traits make signatures easy to forge vs. traits that make them difficult. Signatures were forged to reveal the quality of the initial perceptions. It is expected that the research will provide evidence of mechanical traits that do in fact indicate ease of forgery to a human subject, as well as those traits that contribute to the difficulty of forgery. There are a number of digitizers on the market, and making sure that they interoperate is key in developing a DSV solution. This research will examine which variables are interoperable across a number of digitizers. In conjunction with the work done in SC37, a study is underway which examines the "force" variable to understand whether digitizers on the market are calibrated correctly. The final research is Repeatability of Signatures. In test situations, subjects are asked to sign repeatably, however, it is anticipated that these signatures are not the same as those signed in "real" applications. This study will examine which variables wander when signatures are signed repeatedly. Previous Work and Publications
|
As the use of signatures for identification purposes is pervasive in society and has a long history in biometrics, dynamic signature verification (DSV) could be an answer to authenticating a document signed electronically. Dr. Elliott's research interest is in the field of dynamic signature verification and the testing and evaluating of dynamic signature verification. Research in this area centers around establishing testing protocols for dynamic signature verification; the analysis of DSV variables and their performance across different types of digitizers; as well as contributing to the ISO/IEC JTC1 SC37 work on 19795-7 and 19795-11.