Avoiding fraud in biometric identification

Spanish scientists from Carlos III University of Madrid are analyzing possible attempts at fraud in various biometric identification systems in order to improve the security of facial, iris, fingerprint, or vascular recognition, among other types.

The field that these researchers are working in is known by its nickname, “anti-spoofing,” and basically consists in trying to detect all of the possible attempts at fraud that a biometric system might suffer, especially with regard to an action in which the user presents the biometric proof to the system.

“What we are trying to do is detect those attempts so that the system can then act accordingly,” explains the head of UC3M’s Grupo Universitario de Tecnologías de Identificación (GUTI)(University Identification Technology Group), Raúl Sánchez Reíllo, who is leading this research. This way, if someone used a colored contact lens to recreate a specific iris at an access control point, the system would detect this possible fraud attempt and would indicate to this user that he or she could not use the automatic system and would have to use the manual identification system, with a security agent, for example.

These scientists work on anti-spoofing related to most of the forms of biometric identification. In addition, they evaluate the strength of current biometric systems in the face of various types of attacks, and they also create algorithms, devices and collateral techniques and usage policies that avoid and detect these attempts at fraud. “Currently, we are working very intensely on the ocular iris as well as written signatures, although previously we have worked on fingerprints, and in the near future we will be working on facial recognition,” comments this professor from UC3M’s Electronic Technology Department (Departamento de Tecnología Electrónica de la UC3M), pointing out that the challenges in this field are enormous. The reason: there is a constant struggle between the good guys and the bad guys, in which the latter is constantly trying to find new ways to attack the security of the system. “Let’s say that the good guys work to stay a step ahead of those attempts, introducing anti-fraud measures in advance of what the bad guys might come up with,” he reveals.

At a recent scientific conference, the scientists of the GUTI group presented part of their research on sources of noise and the most common falsifications in recognition systems based on iris identification. The article, titled “Estudio de la Casuística de la