St. Gallen police rely on super-recognizers in the hunt for criminals
Published: Monday, Jul 15th 2024, 09:50
Back to Live Feed
In the fight against criminals, the St. Gallen police are now relying on people with above-average abilities to recognize faces. Super-recognizers sometimes recognize faces more reliably than software. Six of them work for the St. Gallen corps.
He has seen this man before, a police officer from the St. Gallen municipal police is convinced during a domestic violence operation. The operation is still ongoing and the police officer is not sure why he knows the man.
It later transpires that the city police officer recognized a criminal he had seen on surveillance camera images during earlier investigations. His skills as a super-recognizer helped him.
"Super-recognizers are people who have a pronounced ability to recognize faces and can remember faces particularly well," Stefan Helfenberger, Head of Investigation Support at the St. Gallen Cantonal Police, told the Keystone-SDA news agency. Science assumes that only two percent of people have these abilities.
Six police officers recognized as super-recognizers
Helfenberger is leading a pilot project for the cantonal and municipal police of St. Gallen. The police forces are testing the use of super-recognizers. According to scientific tests conducted by the University of Lausanne, six police officers can memorize faces exceptionally well and recognize them. This means that they are able to compare different faces with each other - in principle like facial recognition software - and determine whether they are the same person.
Stefan Helfenberger named image searches as a current focus for the super-recognizers. Specifically: the police have surveillance images of various robberies, for example. Super-Recognizers look at the images, compare them and can use their skills to assess whether the perpetrators are the same.
Or they compare the surveillance images with mugshots and see if they recognize identical people. Depending on whether the persons in the mugshots are already known by name or not. These findings are used in further investigations.
Bed of nails convicts perpetrator
In the episode mentioned at the beginning, the super-recognizer of the St. Gallen city police, who wishes to remain anonymous, primarily recognized the man's face. Even though super-recognizers are primarily specialized in faces and scientific tests exclusively determine the ability to recognize faces, other features can also provide clues.
In the above-mentioned example, the perpetrator's hands could be seen on other images. In these pictures, he noticed that the perpetrator had a special nail bed on one thumb. At the police station, the super-recognizer looked at his hands. He recognized the nail bed. "I was sure I was looking at the person I saw on the CCTV images." Investigations confirmed his suspicions and the man was found guilty of several crimes.
A few seconds are enough
The city police super-recognizer doesn't just pay attention to faces. "I've noticed that I'm good at recognizing a person's posture, gait or gestures," he continued. When he compares images on the computer, it only takes a few seconds to decide whether he has seen a face before. However, the super-recognizer must have consciously perceived a person. "I can't walk through the city of St. Gallen on a Saturday and recognize every face afterwards."
The super-recognizers at the cantonal and municipal police had to prove their skills in a multi-stage scientific test. In one test, they first memorized unfamiliar faces.
Later, they were asked to recognize the faces under difficult conditions: Changed facial expressions, different viewing angles or poorer image quality. The police officers had to decide which of the three faces shown they had already memorized in the learning phase. In another test, they compared pictures of young adults and middle-aged people and had to indicate which pictures showed the same person.
Advantage over software
The tests were carried out by neuroscientist Meike Ramon. Ramon is a professor at the University of Lausanne and a leading expert in the field of super-recognizers. She is convinced of the benefits of these abilities for the police. It is not possible to make general statements about whether facial recognition software or a super-recognizer is more reliable in recognizing faces. But: "Our research shows that there are people who can deduce the identity of unknown persons with very little image material, despite poor image quality and age-related changes, in a way that algorithms cannot."
The pilot project is expected to run until the end of the year. The cantonal and municipal police then want to decide whether the super-recognizers will become a permanent part of police work and whether they will be used in other areas.
©Keystone/SDA
English 
German 
French 
Spanish 
Arabic