In two to three minutes, talented people can match a hundred faces with mugshots – and recognize a wanted person. The Winterthur city police want to take advantage of this.
Super Recognizers are extremely fast and good at comparing unfamiliar faces or memorizing them and recognizing them later.
In many countries, police forces are increasingly using people with special facial recognition skills for manhunts. In Switzerland, the Winterthur city police are leaders in the field of such super-recognizers: Lorenz Wyss, head of the search department, has dealt with super-recognizers more intensively than almost any other Swiss police officer. Wyss told the NZZ that it is a missed opportunity that Super Recognizers have only been called in on a selective basis in Switzerland.
The Winterthur city police have now reacted and, starting in January, will be the first police force in Switzerland to use a super-recognizer in the Investigations and Prevention department. Wyss confirmed this when asked by the NZZ. The employee has previously worked as a police officer in another municipality and was hired by the Winterthur City Police Department because of his special skills. However, it is still too early to say how exactly and in which cases it will be used.
The phenomenon has only been known for around 15 years
Super recognizers are extremely fast and good at comparing unknown faces or memorizing them and comparing them later. You need little experience and information to recognize faces. Especially when the conditions get difficult, they are unbeatable. For example, if the images are blurred. Or if the faces are taken from an unfavorable perspective. Some Super Recognizers excel in one area, while others consistently outperform across the board.
The existence of super-recognizers has only been known for about fifteen years. However, scientists do not yet know exactly why the facial recognition abilities are so different. And because the exact procedure of the brain has still not been decoded, the abilities can only be reproduced artificially to a limited extent. So far, super-recognizers have also been superior to algorithms. For example, you can also recognize people when they are masked or the picture is blurred from a great distance. Digital facial recognition systems are reaching their limits here.
The Winterthur super-recognizer has already been able to solve cases at his previous place of work because he remembered where he had previously seen people from mug shots. He told the NZZ a few weeks ago. A few years ago, the man noticed that he regularly recognized people whom he had previously seen in a completely different situation – and often only for a short time. Tests then found that it recognized faces much faster and more reliably than most people. Today it takes him two or three minutes to find the one out of a hundred images with different faces that matches that of a surveillance camera.
Advantages over facial recognition software
So far, super-recognizers have met with little interest in Switzerland. Various cantons rely on facial recognition software instead. This performs tasks similar to those performed by the police for super-recognizers: if, for example, a bank robber is caught by the surveillance camera during a robbery, this image can be compared with photos of people who were recorded by the police during earlier criminal proceedings. Software and super-recognizers perform these tasks faster and more reliably than is possible with traditional matching.
But when it comes to facial recognition software, the concerns are great. Various NGOs are demanding that politicians stop development and ban automated facial recognition. In a representative poll this year, just a third of respondents said they would allow law enforcement agencies to use voice and facial recognition software.
There are also doubts as to its legality. Various lawyers are of the opinion that the Swiss code of criminal procedure and the cantonal police law do not form a sufficient basis for the use of facial recognition software. The deployment of police officers with special skills, on the other hand, appears to be unproblematic from a legal point of view.
On the rise in Germany
In Germany in particular, more and more police corps are using Super-Recognizers. When there were mass sexual assaults around the train station in Cologne on New Year’s Eve 2015, the Cologne police were the first to get support from their London colleagues. Two super-recoginzers helped the people of Cologne to evaluate the countless cell phone and surveillance images. In the meantime, the Cologne police have formed their own team of Super Recognizers.
The Munich police followed a short time later and identified 37 super-recognizers among 4,500 employees. This year, the Berlin police also asked 18,000 police officers to clarify their facial recognition skills as part of tests. After the attacks on the Christmas market on December 19, 2016, people there would have been happy to have super-recognizers to track down the assassins.