Scanning away, keeping the thieves at bay

Scanning away, keeping the thieves at bay
Excited: Member of Parliament Dr Lily Neo (left, in grey), guest-of-honour at the Singapore International Jewelry Expo, admiring the 50.55-carat diamond. Models (far left) showing off jewellery in a fashion show during the opening ceremony yesterday.

It seemed like a scene from a movie: Stern-looking men huddled around screens scrutinising faces flashed on closed circuit televisions (CCTV).

Suddenly, one of the officers freezes the footage and zooms in on a face on the screen.

Within seconds, photos of known criminals are brought up. The software automatically matches the suspect's face with those on the list, with a percentage number indicating the level of similarity.

It is not a movie. This is being played out daily at the Singapore International Jewelry Expo 2014 (SIJE) at Marina Bay Sands. The expo started yesterday and will end on Sunday.

The facial recognition technology is part of stringent security measures to safeguard the US$200 million (S$250 million) worth of fine jewellery, gemstones and watches on display.

The CCTVs monitor the three entrances to the exhibits.

Besides the usual metal detectors and infrared sensors, the facial recognition system is the most sophisticated security system in place.

It was set up by the Asia-Pacific wing of Axxonsoft Asia, a video surveillance and security solutions company based in Russia.

Every visitor who passes through the three exhibition doors is scanned and checked against an extensive watch list from various sources such as the Singapore Police Force and organisers of previous high-security events.

Mr Durairaj Gireraj, the Asia-Pacific director of Axxonsoft Asia, said: "The system works by scanning the facial features of the people who appear on camera, and is designed to match and scan faces regardless of gender, age or race."

The software recognises suspects even when they try to disguise themselves, including using fake moustaches or sunglasses.

"The system is able to recognise and match anyone on the watch list, short of them undergoing plastic surgery," said Mr Durairaj.

There are three levels of alert: Green, yellow and red.

The green level indicates that none of the faces scanned exceed a 50 per cent match.

A yellow alert occurs when there is a 50 to 80 per cent match in facial features.

Red alerts occur only when the system flags a person with a facial resemblance of more than 80 per cent.

A security team will immediately be dispatched to nab the suspect and hand him over to the police.

Alerts would also automatically be sent to all security personnel.

When The New Paper was in the security room yesterday, a yellow alert was triggered by a woman in her 50s.

She had a 59 per cent match with one of the male suspects on the watch list.

Different people

But a quick glance by security officers confirmed that they were two different people, though there was some resemblance.

Said Mr Steven Tan, the general manager of CEMS-Conference and Exhibition Management Services, the company hosting the exhibition: "This exhibition is different from something like a food fair where vendors can dismiss something like a bag of coffee powder being stolen.

"Diamonds cost so much more and are also easier to hide due to their small size.

"When you are in charge of an exhibition with US$200 million in jewellery present, you will do anything necessary to ensure that everything stays in place."

"With such technology, I am at least able to sleep better," he said with a laugh.

He recalled a theft last October at a similar exhibition, when a S$400,000 ring was stolen at the Singapore Jewel Fest at Ngee Ann City Civic Plaza.

"We have talked about that and analysed the possible lapses of security to ensure that the same thing does not happen at our event.

"A show like this without incident is a good show," Mr Tan said.


This article was first published on JULY 4, 2014.
Get The New Paper for more stories.

More about

Purchase this article for republication.

BRANDINSIDER

SPONSORED

Most Read

Your daily good stuff - AsiaOne stories delivered straight to your inbox
By signing up, you agree to our Privacy policy and Terms and Conditions.