The University has expanded the number of greeter kiosks that are around campus — and the kiosks will catch you maskless in 480p resolution.
Designed by UNC students under the guidance of Steven King, chief innovation officer for the Reese Innovation Lab, the Health Greeter Kiosk uses artificial intelligence and computer vision to detect if people are wearing masks and if they are social distancing.
Though first seen at Kenan Stadium in the fall, the kiosks are now deployed at select locations on campus — including Carroll Hall and the McColl building.
King said the main goal for the project was to figure out how to use technology to encourage people to wear masks and to spread out.
The kiosks work by analyzing video frames for human faces and determining if they are wearing a mask or not. If the kiosk doesn’t detect a mask, King said it will display a positively-worded message encouraging mask-wearing.
“The kiosks look like they’re recording video, but they’re not,” he said. “They take a single frame of video, they analyze it, and then immediately in one-thirtieth of a second, take the next video and analyze it and the next frame and the next frame.”
Natalie Huggins, one of the student user experience designers for the project, said that student reactions were a major discussion when developing the project.
“Some people don’t want to be on camera, and then they’re going to wonder if they’re being recorded or if they’re going to be reported,” she said. “We really had the design in mind to think about those student concerns. Those are human concerns for any kind of human or mirror technology.”
King said that the kiosks do not employ any facial recognition technology and that all data is collected anonymously.
“We don’t have any identification, no personal information,” he said. “All we’re doing is collecting how many people are wearing masks.”
Robert Daigle, an artificial intelligence innovation leader at Lenovo, said that the AI used in the kiosk project was trained beforehand with general data protection regulation-compliant data so that it doesn’t collect or use uniquely identifiable data.
“Whenever an image comes in, it takes all the features in that image to identify if there’s a face mask and if it’s being properly positioned on the face or not, and is able to give real-time feedback through the screen directly to the person,” Daigle said.
While this project involved Lenovo engineers who helped to optimize the kiosk’s technology, Huggins said that it was important that students were heavily involved in the design of the on-campus project.
“The most valuable part of it actually being designed by students is that they’re the ones that are closest to at least the technology that we were building just for the school," Huggins said. “Having the user experience designers being students was important to being in touch with how people are going to use this, how people are going to interact with it.”
Huggins said that the next step is seeing the different possible directions the kiosk could go.
“The current project right now is seeing if this can be used for convention centers or see if it can be used in getting people in a larger group if they want to rent this kind of kiosk,” Huggins said. “But none of that is for sure.”
To get the day's news and headlines in your inbox each morning, sign up for our email newsletters.