Facial recognition systems have been around for longer than many might think. Primitive versions of the technology have existed since the 1980s when mathematicians began defining faces as a series of numerical values and used probability models to find a match. The underlying technology has changed significantly in recent years; the old tech has been replaced by a new generation of algorithms, and they’ve become remarkably effective.
Fuelled by advances in processing power, cloud computing, artificial intelligence and machine learning, algorithms have grown significantly more sophisticated and prices have fallen to commercially viable levels.
Applying automated video analysis to footage from CCTV systems does raise several ethical and legal questions, which have been widely discussed in the media and civil rights groups, but there is no doubt that it has the potential to assist the emergency services to increase public safety in preventing and detecting crime as well as potentially saving lives.
Facial recognition has many potential benefits for modern society. It can be used as a biometric for verification – indeed it is increasingly used in this way. Passports, computer and phone log in, buildings access, even mobile payments, can use face recognition to smooth our way through ’gateways’.
When combined with other capabilities such as analysis of body movements, facial recognition can contribute to the detection of unusual behaviours that might be indicators of ill health such as heart attacks or strokes. Used in public places or other locations like care homes, this could trigger calls for assistance – and ultimately save lives.
In the UK, several trials of the technology have taken place over the last 12 months. Let’s look at some of these now.
The Metropolitan Police
The Metropolitan Police has been trialling Live Facial Recognition technology in the last year and has been testing it in a range of environments including public events and crowded public spaces. There have been ten trials including;
Notting Hill Carnival in 2016 and 2017
Remembrance Day 2017
Port of Hull Docks (assisting Humberside Police) in 2018
Stratford Transport Hub for two days in June and July 2018
Romford town centre in February 2019
A full independent evaluation is now underway and the findings, expected shortly, will help inform how the Metropolitan Police and other forces might use the technology in the future. They will also hold a public consultation to discuss the use of facial recognition and to canvas public concerns about it. An Advisory Consultation Group of key stakeholders has been set up to discuss the use of live facial recognition.
South Wales Police
Ahead of last year's UEFA Champions League Final in Cardiff, South Wales Police revealed that it would be using facial recognition technology to identify potential troublemakers - be they ticket touts, hooligans or terrorists.
The tests used the city’s CCTV network and mobile police vehicles to capture faces at the train station, in the town centre and around the stadium. That data was compared against a watch list, which allowed police to act if the software brought back a match from custody photos.
In this short video Inspector Scott Lloyd of South Wales Police answers questions on the new Automatic Facial Recognition system.
Use Cases for Facial Recognition
1) Smarter Surveillance
CCTV catches the average Londoner about 300 times a day. But with over six million surveillance cameras nationwide – more than in any other European country – it’s not just the capital that’s well equipped to utilise CCTV for facial recognition. Local authorities, as well as the emergency services, across the country can leverage this existing infrastructure as a launchpad for new opportunities.
As technology develops, and the Internet of Things becomes more widespread, the opportunities for cities to make more of surveillance grows. In January 2019 it emerged that local authorities in England have spent more than £750m on CCTV over the past decade, an increase of 17 per cent per year since 2010.
Video surveillance has existed for decades but trawling through hours of recorded footage for a single individual or a moment limits its usefulness. Artificial Intelligence can bypass these limitations and search through surveillance footage for specific patterns at a rate that humans could never meet, potentially identifying threats, emergencies, or even individuals at a moment’s notice.
This technology is already in use today. In August 2017, the New York Department of Motor Vehicles revealed that AI-powered facial recognition technology was used to aid in over 4,000 arrests by utilizing the DMV’s database of driver’s license headshots. Video analysis is getting smarter too; a pilot program in Japan utilizes AI-powered surveillance systems in rail stations to identify anomalous behaviour indicating intoxication or criminal behaviour, heading off more serious incidents before they occur.
High-quality surveillance footage is, of course, foremost a tool for the police. It can be very persuasive, both in and out of the courtroom. Real-time surveillance can help catch offenders red-handed; mobile surveillance can livestream local CCTV images directly to the smartphones of officers on the ground. Or in the air. Smartphones, vehicles, drones and helicopters are all used to stream data and view video. It means decision-makers can work anywhere. Images and videos can form strong evidence in an investigation. And they can help police crackdown on terrorist threats too.
2) Missing Persons
Facial recognition can be used to find missing children and victims of human trafficking. If missing individuals are added to a database, control rooms can become alerted as soon as they are recognised by face recognition—be it an airport, retail store or other public space. In fact, 3,000 missing children were discovered in just four days using face recognition in India.
3) Medical Emergencies
Facial recognition software can also be used to identify unconscious or severely injured individuals at crime scenes or accidents. If an unconscious or injured person was identified and their medical information was linked to their data, lifesaving medical history information could be provided to paramedics so that the patient could be treated earlier.
The potential of emotion recognition, in addition to facial recognition, is already exciting security companies and law enforcement organisations across the globe, due to its ability to determine an individual’s state of mind or intent through their facial cues, posture, gestures and movement. The fact that this can be done from different angles, and even if the subject is moving or partially obscured, say by a balaclava, as well as under various light conditions is particularly impressive.
5) UK Government Strategy and Legislation
In a recent House of Lords debate tabled by Baroness Jones of Moulsecoomb, the Minister of State to the Home Office, Baroness Williams of Trafford, stated that the government will publish the Home Office biometrics strategy in June 2019, and this will address the use of facial recognition technology. She also stated that there is ongoing work to implement last year’s custody images review, which provides a right to request deletion, and that the government are planning improvements to the governance of police use of custody images and facial recognition technology.
The Home Office has published the Surveillance Camera Code of Practice, which sets out the guiding principles for striking a balance between protecting the public and upholding civil liberties. Police forces are obliged under the Protection of Freedoms Act—POFA—to have regard to this code. Similarly, the Information Commissioner’s Office has issued a code of practice, which explains how data protection legislation applies to the use of surveillance cameras and promotes best practice.
Integration of the technologies discussed in this blog will take time to be fully integrated into the control room of the future, but there is no doubt that once the technological, legal and ethical challenges have been worked through, facial recognition will make our streets safer.