Facial recogware in the spotlight at World Gaming Protection Conference next week in Las Vegas

February 7, 2022 11:59 PM
  • Buck Wargo, CDC Gaming Reports
February 7, 2022 11:59 PM
  • Buck Wargo, CDC Gaming Reports

Casinos are expanding their use of facial-recognition technology that experts say has made tremendous strides in identifying unwanted patrons on their gaming floors.

Story continues below

While some technology isn’t as effective in an era of facial masking due to COVID-19, look for casinos to adopt its use more and more in the future, according to Willy Allison, managing director and founder of the World Game Protection Conference that runs Feb. 15-17 at the Tropicana Las Vegas.

It is anticipated that the overall facial-recognition marketplace is expected to grow from $3.8 billion in 2020 to $8.5 billion by 2025 and the casino industry is expected to be part of that.

The conference session will be headed by Patrick Grother, a scientist at the National Institute of Standards in Technology that’s responsible for biometric-algorithm evaluation and biometric-performance-testing standardization. He leads the Face Recognition Vendor Tests, which constitute the world’s largest independent public tests of face-recognition algorithms and provides findings of their research.

“It’s been around for 15 years or so, probably longer, and the product has never worked for casinos because of various factors, such as lighting, plus the camera technology wasn’t as good,” Allison said. “There was no high definition or 4K. Now that we have camera technology, the software behind facial recognition is more accurate. In the last three years, the technology has jumped by leaps and bounds and casinos are not as skeptical as they were. They’re purchasing it and sales over the last three years have gone up.”

Allison said 10 years ago, less than 10% of casinos deployed facial-recognition technology. Today, it’s at least one-third of casinos and growing.

“I see it as the future,” Allison said. “As years click by and more people are barred, it’s impossible to keep up with this. In some jurisdictions, casinos can be fined by regulators for allowing self-excluded people from coming into their properties.”

Allison said those barred people range from criminals to card counters, cheats to money launderers, and those with gambling problems who self-exclude themselves from casinos.

Every casino has its own database of barred players and those lists grow to be large after 20 to 30 years, Allison said. It’s unrealistic to think that any person, especially with turnover, can remember 10,000 faces.

“No matter how much of the staff leaves, it helps you do the job,” Allison said. “It doesn’t confirm someone’s identity. It just runs a match through the pictures of the existing database and increases the odds. You still have to go down there (to the casino floor) and ID them.”

Grother said an “industrial revolution” over the last seven to eight years is continuing, based on improvements with artificial intelligence. It has helped that the number of developers of the technology has increased substantially over the last couple of years.

“The new generation of facial-recognition algorithms is capable of recognizing images in environments and circumstances that weren’t possible a decade ago with adverse lighting and at some distance, with people not looking at the camera,” Grother said. “But while the reliability and capability have improved, some limitations still remain.”

Face masks are a mixed story, with some algorithms not working, but the industry responded and some technology now works with masks, Grother said.

“Face masks put the industry back about two or three years in terms of accuracy,” Grother said. “They are today where they were in 2018 and 2019 for unmasked photos. But face masks are not fatal, compared to where they were pre-pandemic.”

Grother said if people cover up more of the face with baseball hats, sunglasses, and face masks or add disguises, that makes it difficult to identify people. The other limitation is aging, if someone hasn’t been in a casino in 20 years.

“Another problem, for some reason, if someone is in a shadowy environment where there’s not much light on the face and they never look at the camera, that can undermine recognition,” Grother said. “Algorithms still like a frontal view, but the envelope to capture has increased a lot.”

Casino surveillance people have greater assurances and a degree of confidence that the facial-recognition technology works and collects the information with cameras at entrances to their properties, Allison said.

“In the old days when a photo came up, it might say it could be one of these 25 people,” Allison said laughing. “We’ve had issues with gender and demographics and accuracy that civil libertarians were saying was unfair with false negatives coming up.”

Grother recalls his discussions nearly a decade ago with three Las Vegas casinos in which they told him the technology wasn’t good enough at the time. That’s a different conversation today, he said.

“There’s more potential for the use of facial recognition, because the industry has undergone this revolution,” Grother said.

Register for the World Gaming Protection Conference here.