Cancelable Authentication Based On Fingerprints Texture

The challenges during the deployment of biometric systems are numerous. We have to take into account performance, privacy, security and acceptability issues. In this context, we developed an authentication system using fingerprints whose extraction phase of the representative model breaks up into two stages. In the first stage, a vector characteristic of the fingerprint image is derived by using a Gabor filterbank. A detection process of the area of interest is operated beforehand. During the second stage, the characteristic vector is plunged in a salting scheme by using a secret key K. The model thus generated is a binary vector called BioCode. Only the BioCode will be registered for the recognition which is done in a match-on-card system. The BioCode thus obtained offers an elegant solution to the mentioned problems before as our phase of tests proves it.