| Perspiration for Detecting
Liveness in Fingerprint Scanners—Comparison of Different Classifiers |
| |
| Reza Derakhshani, Sujan Parthnasardi, Lawrence Hornak, Stephanie
Schuckers |
| |
| Liveness detection, i.e. determining whether an introduced biometric
is coming from a live source or not, has been suggested as a means
to circumvent attacks that use spoof fingers. An anti-spoofing method
based on liveness detection has been developed in our laboratory for
use in fingerprint scanners. This method quantifies a specific temporal
perspiration pattern present in fingerprints acquired from live claimants.
We are currently optimizing our patent pending perspiration detection
algorithm for different fingerprint scanner technologies using a larger,
more diverse data set and a shorter detection time window. Several
classification methods are tested in order to separate live and spoof
fingerprint images. Each method has a different performance with respect
to each scanner and time window. All the classifiers achieve approximately
90% classification rate for all scanners, using the reduced time window
and the more comprehensive training and test sets. Based on the classification
results, we believe that this perspiration-based liveness detection
method has the potential to reduce the susceptibility of fingerprint
scanners to spoof attacks. |
| |
   |
| |
| Figure: Fingerprint images show a distinctive moisture pattern that changes over time related
to perspiration which can be used as a measure of liveness for fingerprint recognition. |
| |
| To be published in IEEE Transactions on Systems, Man, and Cybernetics,
Part C: Applications and Reviews |
