Computing Power and Recognition Technology Come to Brick-and-Mortar Stores

By John Blyler, Editor-in-Chief

Embedded technology is set to enhance the productivity of retailers, advertisers, and shoppers alike. Ryan Parker, director of marketing for the Intel® Embedded & Communications Group, talks with Embedded Intel® Solutions (EIS) magazine about the technology behind the trends.

John Blyler: What lies ahead for Intel in the embedded space?

Ryan: While multicore and virtualization are among our more advanced embedded technologies, we are also seeing a lot of interest in manageability and security solutions. By manageability, I mean the remote management of assets that are not located close to your IT shop ranging from PCs to globally distributed embedded systems—not only remotely managed via software applications, which has been done for a long time now, but that turn the entire system on and off.

Such management applications also are used to save power. Manageability means you can turn off a device and then turn it back on remotely. This can be done to repair a down system and to save energy, which means a real cost savings. For example, you would actually turn the entire motherboard down—especially if it is powered over USB. Now, all of a sudden, with your sleep states and everything else, you can power the system down and turn off all the peripherals.

JB: Power management is important. But is it the main driver?

RP: One of the biggest trends is in productivity-enhancement technology. For example, look at the retail space, where they’re using 15-to-20-year-old technology in point-of-sale (PoS) systems, paper advertisements, and changing rooms with no interactivity. Retailers want to update their systems to bring the online experience into the brick-and-mortar establishments.

Here’s a scenario: Let’s say that I want to look for a new pair of pants, but I don’t want a high-pressure salesperson telling me what to buy. Instead, I would go to a kiosk to locate the clothing [see this issue: “Intel Consumer Technology Update”]. This kiosk will also cross-reference my search to suggest other things that I might like to buy. The clothing would have a small radio-frequency-identification (RFID) tag so that I could scan the items in, receive an e-receipt on my cell phone, and leave. That way, I’m in and out of the store without needing a lot of personal interaction. More importantly—to the retailer—I didn’t just leave because I got frustrated and couldn’t find what I wanted.

Another area of growing retail application is in digital signage. Flat-panel displays with an embedded processor are getting cheap enough to replace paperbased signs. We are seeing more and more of these displays in the U.S., but some unique opportunities are presenting themselves in the Japanese subway. People use these subways in very routine ways: Children go to school, workers go to work, and the shoppers go shopping. Their routes are very regimented, occurring at the same time each day. Knowing this audience, if I could update my signs with the content that they care about at that specific moment, my advertising dollars would probably triple in value compared to static paper signs. I have THE sign that I need at the exact time that I need it.

JB: How do you obtain the demographics to know—in real time—the audience that your signage might attract?

RP: There are a couple of ways to obtain that information while avoiding privacy issues. One way is to sample the audience, as is typically done. Most of this demographic information already exists, such as what kind of people walk around Times Square in New York City at any given time of day. The other approach would provide information on gender, age, and other kinds of demographic information by integrating a camera-based system.

JB: Facial recognition?

RP: We are trying to steer clear of facial-recognition technology— at least for now—because of the privacy issues. Instead, we use imaging technology that estimates the age and gender of a potential retail customer. This data would then be used to target the content on the digital signage, making it more useful to the viewer as well as the advertiser. Recently, we demonstrated such a system at the Intel® Developer’s Forum, where it achieved 90% accuracy in demographic recognition. The system consisted of a simple web camera with appropriate third-party algorithms running on an Intel® processor.

Now, if you layer the demographic-recognition system with the manageability technology that I mentioned earlier, you’ll achieve quite a productivity enhancement. The recognition system will help to target retailers’ advertisement dollars where they do the most good. Meanwhile, the management system will help to reduce system errors remotely so customers don’t see the infamous blue-screen error on the display. But all of this requires a considerable amount of computing power—that is, to run the data-intensive recognition software and the display video streams.

Another benefit with a camera-based display system is the power savings and added security over existing laser-based scanners. Laser scanners draw a lot of power. They are not well adapted to turning on and off. Conversely, a camera system uses considerably less power. In addition, a camera can not only read a barcode; it can also “look” at the size and shape of the object being scanned to make sure they match up. This will help to reduce fraud—namely, someone placing a 99-cent barcode on a $1000 bottle of wine.

JB: Thanks for sharing all of these interesting applications.

Ryan Parker is the Director of Marketing for the Intel® Embedded & Communications Group. Ryan has a BS in Industrial Engineering from Montana State University and an MBA from Arizona State University. He started with Intel in 1996 in engineering and has worked in all aspects of the semiconductor business including product development, manufacturing, sales, and marketing.



John Blyler is the Editorial Director of Embedded Intel® Solutions magazine. John can be reached at: jblyler@extensionmedia.com