Fraunhofer Institute for Computer Graphics Research IGDThe suitability of biometric samples for automated identification is not the same for every sample. External and internal factors can influence the biometric quality. It is therefore desirable to have the ability to judge the biometric quality of a sample in order to set minimum requirements or to trigger a new biometric extraction and thereby avoid subsequent errors during recognition. In this project, we are developing algorithms for assessing the biometric quality of facial images and other biometric modalities.
Among the areas we analyzed in the course of our research were the regions of the face that have relative influence on the overall suitability of an image for facial recognition and thus on the biometric quality. The results showed the importance of the eye region for biometric face recognition, in contrast to the less important regions of nose and mouth.
An important but perhaps not obvious fact regarding biometric suitability of images for biometric recognition is that it does not necessarily correlate to human assessment of quality. In a recent study, we investigated how special algorithms for the determination of biometric quality compare with algorithms for general image quality. The experiments we conducted showed that there is a correlation between image quality and biometric quality, but that methods specifically designed to determine biometric quality perform much better than algorithms for image quality determination.
The research on Biometric Sample Quality is part of the Next Generation Biometrics Systems mission within the ATHENE project. ATHENE, the National Research Center for Applied Cybersecurity, is funded by the German Federal Ministry of Education and Research (BMBF) and the Hessen Ministry of Science and the Arts (HMWK).