New technology allows retailers to spot a celebrity approaching
It is a moment of pure panic for a shop assistant - you recognise the person in front of you demanding attention but you cant place them do not have time to deal with them.
Then, as soon as they leave the shop, the penny drops and you realise you have snubbed one of the rich and famous, and your boss is not going to be happy that you have turned away a month’s worth of sales or a glowing celebrity endorsement.
But now a purpose-built facial-recognition system has been designed to ensure no hapless shop assistant accidentally snubs their best customer again, the Sunday Times reported.
The VIP-identification technology, created by NEC IT Solutions, is already being tested in about a dozen top stores and exclusive hotels in Britain, America and the Far East.
The Cambridge based company believe it could help provide a personalised service and improve customer service.
It could prevent blushes such as those of watersports store owner Dave Buckland, who failed to recognise the Duchess of Cambridge when she went in to his shop to buy a wetsuit last year.
The company already supplies software to help security services spot terrorists and other criminals, works by analysing footage of customers' faces as they walk through the door.
Various measurements of the face are taken to create a numerical code, known as a face template, which is checked against a database of clients.
An alert is sent to staff via computer, iPad or smartphone, providing details that might include their names, dress size, favourite room or previous spending.
Chris de Silva, a vice-president at NEC IT Solutions, said: " The luxury end of the market is quite interested in it — they're interested in VIPs."
Because they use the neural network to create a holistic view of the face, it still works when people wear sunglasses, hats and scarves in an effort to prevent identification, he claimed, dismissing the idea that it could create privacy concerns.
Recent tests had found that facial hair, changes in weight or hair colour and the passage of time did not affect the accuracy of the system.
"Essentially if a human can identify a face, our system can also do so — but it never gets tired or bored," he said.
Mr De Silva described the recognition software as an extension of the loyalty-card system, and said it may tempt customers back to the high street by bringing back the personal element.