EAB webinar looks at facial recognition for criminal investigations in EU countries
A recent webinar by the European Association for Biometrics (EAB) explored the role of facial recognition technologies in criminal investigations across the EU Member States.
The talk particularly focused on the analysis of the results of the EU-funded ‘Towards the European Level Exchange of Facial Images’ (TELEFI) project.
The biometric initiative saw the collaboration of a number of European institutions, including the Estonian Forensic Science Institute, the Estonian Police and Border Guard Board, and the Finnish National Bureau of Investigation.
Also, the Latvian State Forensic Science Bureau, the Netherlands Forensic Institute, and the Swedish National Forensic Centre.
The TELEFI webinar featured Professor Dr. Didier Meuwly, principal scientist at The Netherlands Forensic Institute (NFI) and chair of Forensic Biometrics at the University of Twente, and Andra Sirgmets from the Estonian Forensic Science Institute.
Expanding biometric options for criminal investigation
According to existing European regulations, DNA profiles, fingerprint biometrics, and vehicle registration data are currently exchanged between member states through the Prüm framework.
These guidelines were established in 2005, and have been used by member states to tackle international crime, terrorism, and illegal migration.
The Prüm system has been evolving in the past decade and it has reached a point where the EU is considering the introduction of new modalities, including facial recognition.
In this light, the TELEFI project explored the administrative and law enforcement use of facial recognition in criminal investigation, provided a summary of the project, and presented the recommendations for the EU Member States.
Exploring database and legal framing issues
According to Sirgmets, as of 2020, there were 11 EU states where facial recognition for criminal investigation was implemented, with seven additional ones scheduled to complete implementation by the end of 2021.
In these countries, the technology was mainly used to identify suspects via the extraction of images from CCTVs and their analysis via a facial recognition system.
The Forensic Expert then explored how databases of images were and are selected across the EU Member States.
She clarified that these databases mainly stored images of suspects and convicted individuals, and not the general public (though they may include photos of missing persons, for instance).
Despite being used in several EU Member States, Sirgmets added, automated facial recognition currently has a dedicated legal framework only in Hungary. In all other EU countries, the analysis of facial images using automated systems is allowed through several individual laws and implementing acts.
TELEFI Project: Recommendations
The webinar then proceeded with Meuwly taking the floor, and exploring the recommendations based on the results of the TELEFI Project.
The first one, Meuwly explained, is about harmonization of the facial recognition field at a European level, and following six criteria.
These are respectively agreement on standard and quality requirements for control images, on face recognition search methodology, on training human operators on proficiency tests for identification search, and on reporting and common terminology.
The second recommendation resulting from the TELEFI Project, according to Meuwly, is that promoting interoperability of both civilian and criminal processes related to facial recognition would benefit from harmonization in different terms.
In this regard, he particularly named facial image quality standards, database management, manual face comparative methods, and candidate list review methodology.
The third set of recommendations Meuwly mentioned regard the establishment of a forensic working group for the coordination of the harmonization process, while the fourth one regards applying quality recommendations related to quality standards for face image capture methodology, quality, and methodology for face searches.
The fifth one regards the establishment of a large database of facial images for future work in the criminal field.
Additional recommendations ensuing from the TELEFI Project regard the introduction of facial images within the Prüm system, but only after testing the database using decentralized tools.
At the end of the TELEFI project, a report will be published and a conference will be held to share the initiative’s results in order to stimulate discussion on the implementation of facial recognition for criminal investigation across EU Member States.