Multibiometrics authentication in pos application

In traditional Point of Sales (POS) device, the legitimate user is mainly authenticated by PIN (Personal
Identification Number). PIN method has drawbacks like easy to be forgotten. Fingerprint technology,
as the most viable biometrics in the special and restricted hardware environment of POS device, has
been used to increase user convenience. However, due to inherent problems, fingerprint alone can
hardly replace PIN in many high secure applications. Although keystroke pattern as one kind of
biometrics is in research for years, no real satisfied results have been got due to its inherent great
variability. In this paper, we proposed an innovative method by adapting and increase keystroke
pattern technology to POS device and applications. Together with fingerprint system, the adapted
keystroke pattern recognition system constitute a multibiometric system, which helps achieve an
increase in performance, and provide anti-spoofing measures by making it difficult for an intruder to
spoof multiple biometric traits simultaneously. Since this approach is a combination of partially featurebased
and partially knowledge-based, it can be a more deployable alternative to strengthen existing
PIN and pure biometric authentication methods. Experimental results support that this approach can
be better trade-off of the security and user convenience.