SECURITY OF FINGERPRINT SCANNERS

The fingerprint is certainly one of the most well-known biometric characteristics. Unfortunately, fingerprints are often associated solely with their use in law enforcement, which has negatively burdened there.

A fingerprint is the print of the so-called epidermal ridges on the finger tip. The endings and bifurcations of these ridges are identified as minutiae points. The pattern that arises from these minutiae points is a randotypical feature and can be assigned to an individual. Even identical twins have different fingerprints. For the purposes of comparison in law enforcement, as well as in biometrics, not only the fine features (minutiae points) are used, but also the pore structure and the so-called rough features.

The following section provides details on the reliability of a system’s security by means of a fingerprint scanner, that is, how good the FAR and the FRR actually are when the attempt is made to penetrate the system with a dummy. This is meant to show if it is possible, and if so, how difficult it is to fool a fingerprint scanner. For this purpose, tests with various scanners were made. The goal was to outwit the scanner with prepared fingerprint dummies in order to gain access to the system protected by the scanner. In contrast to the already existing articles such as. We precisely document the materials and instruments used in the tests. This will facilitate readers’ reconstruction of the results, as well as a possible reproduction of the tests or individual steps therein.

According to the literature reviewed in this paper, it initially appeared relatively easy to deceive a fingerprint scanner. Articles such as the one by Chaos Computer Club (2004) gave the impression that a biometric system could be tricked in no time at all, and required only the simplest materials to pull off the deception. Other studies (e.g., Matsumoto et al., 2002) make similar claims and describe the ease with which a “dummy” finger can be made and used to fool a fingerprint scanner.

A goal of our research was to not only review existing attacks but to reproduce them and to document the results more precisely than the original papers. Thus, while all essential points are investigated, we could not reproduce some of the results. The experiments also showed that fingerprints are certainly not forgery-proof. They can most certainly be copied. Thus, it is relatively easy, and was also achieved in these experiments, to produce a functional fingerprint dummy. If an attempt is made to copy the fingerprint of a third party without that person’s consent the success rates are much lower compared to cooperating victims. In this context, it is generally assumed that an object of some kind is available that provides a perfect print. In reality, however, many factors influence the quality of the print. For example, if possible, only this person should have touched the object, otherwise it is uncertain whose fingerprint is being processed. Also, the person should not have touched the object too many times or else the prints will overlap or smudge. The object’s surface also plays a certain role. Furthermore, it must be kept in mind that if possible, the object to be used should be portable. If one takes a fingerprint from, e.g., a door handle, it is rarely possible to do so without being noticed or leaving behind traces.