Shopping Centers Today is the news magazine of the International Council of Shopping Centers (ICSC)
Issue link: https://sct.epubxp.com/i/62026
but it will identify them in broad terms — male, female, older, younger — and show them advertising targeted accordingly. In recent months the likes of The New York Times and ABC News have pointed out the similarities between that movie and real life. "There's probably not a day that goes by that we don't hear this film reference to what we do," said Jason Sosa, founder of Immersive Labs, a face-recognition software company. Sosa was address- ing an audience at an event organized by TED (Technology Entertainment and Design) in November. In reality, the gee-whiz technology has been around for almost as many years as the movie, long enough to de- velop into a billion-dollar market for the hardware and software companies pioneering it. Only in recent years, however, has the technology gained traction, as giants like NEC Corp. and Intel Corp. began to focus on it, along with smaller, pioneering vendors such as TruMedia Technologies and Im- mersive Labs. Digital signage using face-recog- nition technology — known in the industry as anonymous video analyt- ics, or AVA — first gained popularity in Japan, where NEC first tested it in 2008. A year later NEC began selling the system there, and this past Octo- ber the company announced that it would provide the digital signage for two New York City stores of Japanese retailer Uniqlo. The technology is al- ready in use at Uniqlo stores through- out Asia. Standard digital signage has been widely used as an advertising me- dium in shopping centers for years, but face recognition adds the ability to target those ads and thereby in- crease their effectiveness. A camera in the face-recognition video display scans the area nearby and sends data to a computer, where the signal is 170 SCT / MAY 2012 analyzed to detect whether it con- tains patterns of human faces. Those patterns are then compared to facial characteristics such as bone structure and the distance between mouth and eyes that have been "learned" by the software; if the incoming signal cor- responds to a known facial pattern, it is categorized by age and gender and a targeted ad is displayed. The data are also used to create reports that shed light on the overall compo- sition of the market. If the scene in Minority Report helps illustrate how face-recognition technology might someday work in a retail setting, it also raises the unset- tling prospect of the consumer being electronically stalked while strolling through the mall, and this troubles privacy advocates. Indeed, the use of face-recognition technology in gen- eral, especially by law enforcement agencies, has drawn opposition from such groups as the American Civil Liberties Union, which has said that the technology "carries the danger that its use will evolve into a wide- spread tool for spying on American citizens as they move about in public places." To counter such opposition, ven- dors of the technology have adopted the word "anonymous" in describ- ing the technology's marketing uses. It does not record the scanned faces, they insist, nor attempt to identify them through comparison to any cus- tomer database. "No images are ever and will ever be stored for use, review or sharing with any private or govern- mental body," writes TruMedia in its privacy policy message. In developing its AVA offering, called the Audience Impression Met- rics Suite, Intel "wanted to proactively address potential privacy concerns," wrote Jose Avalos, Intel's retail sec- tor general manager, in a September 2011 column on the Digital Signage Today website. "Thus, our technology doesn't recognize individual people, it detects their presence, their dwell time and non-identifying characteristics such as gender, age bracket and height. "Our technology doesn't recognize individual people, it detects their presence, their dwell time and characteristics such as gender, age bracket and height." I wanted people to better understand this difference between detection and recognition, so I insisted on including the word 'anonymous' whenever any- one at Intel discussed video analytics." Ann Cavoukian, commissioner of information and privacy for the Cana- dian province of Ontario, wrote in an April 2011 white paper that AVA "is incorrectly perceived at times to be in the same technology family as surveil- lance and biometrics that can recog- nize, log and track individuals." Mar- ket research in the U.S. has revealed people's objections to marketing that targets consumers based on their be- havior patterns, notes Cavoukian. Fur- ther, she says, privacy advocates warn that a "slippery slope" situation could exist in which evolving technology would later threaten personal privacy in ways it does not now. "Facial-recognition systems and other consumer technologies such as smartphones and RFID embedded smart cards could, in theory, tailor marketing to the purchasing patterns or other characteristics of recognized individuals," wrote Cavoukian. "AVA