Posted on November 7, 2016 by staff

How airports are using facial recognition technology


Liverpool-based Human Recognition Systems has been using automated identification technology since 2001.

The challenge at that time was to take it from being a gadget to a tool, says founder and chief executive Neil Norman.

Today its facial recognition technology, MFlow, is used to track passenger movements at Gatwick, Heathrow, London City and Edinburgh airports, as well as a number of international ones.

Cameras are set up at key locations around the airports – the flow routes – and can track faces from a distance of 8ft to record how long people spend in any area.

“If the cameras have tracked you coming in from the car park, dropping your bag off and then going though security, we can then piece together that journey and know how long it has taken you to go through different flow paths,” Norman says.

This could be of benefit to airlines, he adds, if passengers miss their flights.

“Airports have obligations to the airlines and get fined if people miss flights, but if they have data to say it’s taking 27 minutes to get through security they can respond by opening another lane to get people through.

“At Gatwick we can process 40,000 passengers a day and when you’re looking at numbers like that it’s worth investing in the technology.”

While this sort of tracking was once performed using passengers’ mobile signal, facial recognition has proven much more accurate, Norman says, and is used in most airports.

MFlow works in three stages: acquisition of the image; normalisation, which isolates the specific features that need to be measured; and encoding, turning those measurements into a number to enable any face to be searchable in a database.

Yet, Norman says, the technology never reveals identities and merely uses faces to map journeys and allow the airport to function more smoothly. He accepts there are some reservations over privacy.

“The reason governments lean in the way of facial recognition is that there’s a rich database already in existence, because most citizens have drivers’ licences or passports, whereas to collect the data for iris recognition or fingerprinting would be a major exercise and expense,” says Norman, whose business also offers fingerprinting systems to allow construction workers to access building sites, which was used in the building of London’s Olympic Park.

The technology means you can limit who can use certain types of machinery and access different areas, and allows time tracking.

In the field of facial recognition, Norman says the technology is beginning to plateau and he expects a shift to occur in the future as people seek to evade it.

Filter apps like Snapchat and MSQRD use facial recognition but where the current trend is for people to put their lives on social media, he thinks this will reverse.

“We’re going to see more processing power for the technology with the introduction of quantum computing and AI meaning we can process reams of data faster,” he says.

“But I also think that in the next five years we’re going to see a massive flurry of privacy control technology that will help counter this growth of what are quite simple ways of accessing the rich material that is online.”