Evaluation of Signature Resistance to Attacks
In BSEC’2009 (BioSecure Signature Evaluation Campaign), online signature verification systems were evaluated in terms of resistance to time variability, resistance to change of platform (fixed vs. mobile platform) and finally in terms of quality of the genuine signatures. Such quality was measured by means of a Personal Entropy measure introduced in .
Recent research in the field of Biometrics has shown an increased interest on systems' resistance to attacks. In the online signature modality, verification performance has been evaluated for a long time on databases containing skilled forgeries, which can be of different qualities. This quality can greatly vary depending on the capacity of an impostor to forge a genuine signature, on the difficulty to reproduce such genuine signature, and also on the available information to the impostor about the target genuine signature (dynamic or static information).
In this context, IJCB 2011 Evaluation of Signature Resistance to Attacks (ESRA’2011) aims at assessing the resistance of online signature verification systems to different quality-based categories of skilled forgeries. To this aim, the quality of skilled forgeries is quantified by means of a new measure based on the concept of Personal Entropy, introduced in , relatively to the target genuine signatures.
Evaluation will be carried out on the two largest existing databases containing the same persons, acquired on both a digitizing tablet and a Personal Digital Assistant, namely on BioSecure Signature Corpus DS2 and DS3 respectively. Such two corpuses will allow involving two different protocols of skilled forgeries acquisition.
 S. Garcia-Salicetti, N. Houmani, B. Dorizzi, "A Novel Criterion for Writer Enrolment based on a Time-Normalized Signature Sample Entropy Measure", EURASIP Journal on Advances in Signal Processing, Article ID 964746, 12 pages, doi: 10.1155/2009/964746, 2009.
BioSecure Signature Evaluation Campaign 2011 has the following scope:
- Evaluating the impact of skilled forgery quality on algorithms’ performance, by using the two largest online signature databases containing the same persons: the BioSecure Signature Databases DS2 and DS3. These two databases allow evaluating the impact of mobile acquisition conditions and of the skilled forgery acquisition protocol on algorithms’ performance. The development database released to participants will contain skilled forgery samples labelled according to their quality, measured by means of Personal Entropy .
- Evaluating the impact of different time functions on the resistance of signatures to attacks.
The results of the evaluation will be presented at the upcoming International Joint Conference on Biometrics 2011 (IJCB'2011) to be held from October 11th to October 13th in Washington, USA.
This webiste is hosted by TELECOM & Management SudParis servers.