The research focuses on a new approach to optimising biometric recognition systems. The researchers have developed a method of training algorithms using artificially generated biometric data from children's faces. This innovative approach enables much more accurate identification of real children's faces in large amounts of data.

In turn, the generation of synthetic data is used in other ATHENE research projects in the Next Generation Biometrics Systems (NGBS) research area.

Extended applications

The technology developed goes beyond facial recognition. The team has also created synthetic datasets of adult hands in various poses and gestures. This can significantly improve the recognition rate of hand tattoos, for example, which are often seen in child pornography material.

The research also opens up new possibilities in the search for missing children: The algorithms developed allow for a more accurate age simulation of children's faces, making it possible to make more realistic predictions about the current appearance of a child who has been missing for some time.

The research was carried out as part of the Bio4ensics project funded by the Hessian Ministry of the Interior, Security and Homeland Defence (HMdI). The methods developed for the generation of synthetic data will in turn be used in other ATHENE research projects in the NGBS research area, which is co-ordinated by Prof. Christian Rathgeb and Prof. Christoph Busch, both from Darmstadt University of Applied Sciences, and Dr. Naser Damer from Fraunhofer IGD.

Read more in the press release from Darmstadt University of Applied Sciences