The Emerging of Signature Recognition as a Reliable Biometric Authentication Technology

• Part 3: The Implementation of Signature Recognition –

Penulis: Fajar Kusumaningayu, I Nyoman Aditya Yudiswara, Fadhillah Moulita Andiani

Pembimbing dan Editor: Dr. Eng. Nico Surantha

Figure 1 Enrollment Process

Figure 2 Online Signature Process

Signature recognition has become one of the alternative combinations with Fuzzy Vault to protect the Cryptography keys. Utilizing signature verification before the implementation of the fuzzy vault, thus it set the limit of upper acceptance error rate of a fuzzy vault to the signature validation. The system required the users to prove their self as an authorized person using their signature if they are proven right those signatures will undergo the Fuzzy Vault decoder process to unlock the cryptography key. The study claimed that by using signature verification increase decoding accuracy of around 35% [2].

Conclusion

There are two methods of automated signature recognition which are online and offline signature recognition. The differences between these two methods are in the way to get the signature data. Some of the parameters used in this technique are pressure, velocity, acceleration, position, grid, pixel, etc. In some of the previous studies, signature recognition shows a high accuracy testing result thus making it a reliable authentication technique. It had been used as a login system in other studies as well as became the alternative technique to increase the decoding accuracy of Fuzzy Vault.

Future Work

Previous studies have put their concern on how to get the highest accuracy with minimal error in order to differentiate between genuine and forgeries signature in online and offline signature recognition. Apart from that, the researches done in offline signature recognition have set their goal to provide robust authentication and break down the general idea that online recognition works better.

Despite all of those things, signature recognition faces some difficulty caused by the subject itself. Since the signature is one of behavior biometric authentication, their signature tends to change slightly difference in certain time and condition. Thus it is important to set the accepted range of error where it will not make it fall from the initial purposes of provided authentication but still be able to accommodate inconsistency of people signature.

References

[1] G. M. D’silva and V. A. Bharadi, “Modified online signature recognition using software as a service (SaaS) model on public cloud,” in 2015 International Conference on Information Processing (ICIP), 2015, pp. 360–365.

[2] G. S. Eskander, R. Sabourin, and E. Granger, “Improving Signature-Based Biometric Cryptosystems Using Cascaded Signature Verification – Fuzzy Vault ( SV-FV ) Approach,” in 2014 14th International Conference on Frontiers in Handwriting Recognition Improving, 2014.