Led by intelligence, we will place our effort where it is likely to have the greatest effect.

Live Facial Recognition (LFR) enables us to be more focussed in our approach to tackle crime, including robbery and violence against women and girls.

Today independent ground-breaking research into the Met’s deployment of facial recognition has been published by The National Physical Laboratory.

The study, ‘Facial Recognition Technology in Law Enforcement’ tested the accuracy, in operational conditions, of the algorithm used by the Met in terms of different demographics.

We now better understand the demographic performance of our LFR system. At the setting we have been using it, there is no statistically significant bias in relation to race and gender and the chance of a false match is just 1 in 6000 people who pass the camera.

The findings conclude:

• There are settings our algorithm can be operated at where there is no statistical significance between demographic performance.
• ACCURACY: When used at a threshold setting of 0.6 or above, correct matches (True Positive Identification Rate) were 89%. The incorrect match rate (False Positive Identification Rate) was 0.017%. The chance of a false match therefore, is just 1 in 6000 people walking past the camera.
• DEMOGRAPHIC DIFFERENCE: When used at a threshold setting of 0.6 or above, any differences in matches across groups were not statistically significant. This means performance was the same across race and gender.
• Retrospective Facial Recognition: The true positive identification rate for high quality images was 100%. This is the best performance possible.

Lindsey Chiswick, Director of Intelligence for the Met said: “Live Facial Recognition technology is a precise community crime fighting tool. Led by intelligence, we place our effort where it is likely to have the greatest effect. It enables us to be more focused in our approach to tackle crime, including robbery and violence against women and girls.

“This is a significant report for policing, as it is the first time we have had independent scientific evidence to advise us on the accuracy and any demographic differences of our Facial Recognition Technology.

“We commissioned the work so we could get a better understanding of our facial recognition technology, and this scientific analysis has given us a greater insight into its performance for future deployments.

“We know that at the setting we have been using it, the performance is the same across race and gender and the chance of a false match is just 1 in 6000 people who pass the camera. All matches are manually reviewed by an officer. If the officer thinks it is a match, a conversation will follow to check.

“The study was large enough to ensure any demographic differences would be seen. However, he has also been able to extrapolate these figures to reflect results more representative of watch list size for previous LFR deployments.”

Lindsey Chiswick continued: “We understand the concerns raised by some groups and individuals about emerging technology and the potential for bias. We have listened to these voices. This research means we better understand the performance of our algorithm. We understand how we can operate to ensure the performance across race and gender is equal.”

Link to report: https://science.police.uk/research/resources/operational-testing-of-facial-recognition-technology/