Facial recognition in football stadiums
Recently Danish football club Brøndby announced that it received approval from authorities to start using facial recognition technology in and around its stadium. The news was picked up and covered by many media worldwide (like PlanetBiometrics.com). The club made it clear on its website that it is only using the technology for safety purposes. Facial recognition is used to monitor crowds moving in and detecting people that are on a ban list. Detecting these individuals in an average crowd of 14,000 every home game has proven to be difficult.
Digital rights organizations are skeptical about the use of facial recognition for this purpose. Earlier we reported about the ban of facial recognition by San Francisco and other US cities. In this case the European Digital Rights (EDRi) association is questioning the validity of the approval of this application by the Danish Data Protection Authority (DPA): “The logic of that decision is rather difficult to understand in the present case.“
In Holland academic start-up company 20Face claims to have a solution to apply facial recognition technology in a GDPR compliant way. Their suggested application is not necessarily to detect people on a ban list. They suggest to use it for swift and user-friendly access control, where visitors do the enrollment themselves. They claim to have built “an ecosystem of self enrollment, consent and personal data release management (PDRM) around the core technology of biometric recognition”. The system was tested at a pilot in the Heracles football stadium.
In the discussion about privacy friendly facial recognition it probably makes sense to distinct these two types of applications:
- Scanning a group of people to identify individuals that are on a white list.
- Servicing individuals that have given consent to be enrolled into a system.
In the latter case enrollment and use of the biometrics based application is under control of the individual, which basically means that it is a security system like any other security system that is using cards, biometrics, passwords or a PIN code to identify individual users. On SecIndGroup.com you can see a great application of facial recognition in educational organizations by Panasonic.
In the first case the question probably is whether a normal video surveillance systems is very different from system that is equipped with facial recognition. CCTV-systems are watched by people. With automated face recognition systems (AFR) the guards are helped to do the scanning more quickly. The issue under debate is whether security objectives justify the digital processing of faces of innocent passers-by. We will continue to follow the debate as it further develops.