DSE Day One: Audience Measurement Stands Out

By Travis K. Kurcher
February 28, 2008 - Las Vegas, NV, USA

It's one thing for a retailer to hang a digital sign in a store lobby to inform every consumer who walks in the door about the latest line of fabric softener. But it's quite another to deploy digital signage that recognizes who is looking at the display and target advertising specifically to that individual, based on age and gender.

That's the solution that Capital Networks Limited, a provider of advanced multimedia platforms for digital signage, and CognoVision Solutions Inc., a provider of real-time audience measurement and targeting solution for the out-of-home advertising industry, have partnered to create.

Faizal Javer, the director of operations for CognoVision, explained the solution at the companies' joint booth at the Digital Signage Expo in Las Vegas.

"In a retail environment, a lot of store owners, mall owners and so on, target ads or sell ads to the public," Javer said. "What we are able to do is capture audience metrics. For example, how many people are looking at an ad? We can go as far as age, gender and ethnicity. We can also target ads in real time based on the demographic in front of an ad. For example, if there are three males in a certain age bracket, let's show a relevant ad to them."

A tiny camera mounted on the top of the digital display is able to obtain real-time video of the individuals looking at the screen. Javer says the system uses special software that can determine the number and demographics of the people viewing the ad.

"It has been trained to find certain things," he said. "It knows what a face generally looks like, with very high accuracy. Using that information, we can find things such as gender and find things such as what is a face and what isn't a face. So it generally knows, to a degree, what's not a face, so you don't get things that look similar to a face showing up as well."

A nearby monitor at the booth showed expo attendees what the camera sees. The monitor displayed real-time black-and-white video of the people viewing the display, with colored circles around individual faces.

"This is what our system is seeing," Javer said. "There are two people. They've identified us by green circles just to show that they've found us. Now that the faces are found, we can run secondary algorithms to find gender, age and so on, which can be used to target information as well."

The benefit to deployers, Javer says, is that they're displaying ads in a much more targeted, customized way.

"You can generate more advertising revenue," he said.