ANANAS - Anomaly detection to prevent attacks on facial recognition system

The goal of the project is to develop measures that protect against morphing attacks or quickly discover them. In the area of ​​prevention, this means analyzing weaknesses and possible threats in order to work out specific solution concepts for the various applications.

Abuse detection requires methods that detect, for example, image anomalies that occur in digital imaging in morphing processes. Therefore, within the framework of the project, the images with regard to image processing and camera technology as well as the biometric properties are to be analyzed.

The combination of methods of media and visual forensics promises useful tools for assessing the trustworthiness of face images for identity documents.

Furthermore, an improvement in the field of sensor forensics and image manipulation forensics with regard to the evaluation of authenticity and integrity of image data is sought. With the help of these research results, a demonstrator software for the detection of image manipulations (identity documents) is to be implemented within the scope of the project.