Since schools have sat mostly vacant since last March, when the pandemic forced students to learn from home virtually, video analytics company ZeroEyes has had time to apply its AI to a handful of other situations.
The veteran-run company founded in 2018 initially had school settings in mind when it created its artificial intelligence tool that predicts potential threats with a goal of preventing mass shooter situations from escalating. An average response time for a mass shooter situation in 2019 was between 12 and 15 minutes, according to the company, and its technology aims to curb that by detecting weapons in real time.
“The cofounders were just sick of seeing headlines where kids got killed or people in an office building got killed,” Kieran Carroll, ZeroEyes’ VP of enterprise strategy and government affairs, told us last February. “When we were in the military, we would have loved to have a tool like this.”
Since we checked in with the company a year ago, it’s raised a $5 million seed round and has applied its technology to a host of new environments. ZeroEyes’ platform is now being used in retail locations, corporate campuses, malls and casinos as many schools remain closed for in-person instruction.
While mass shootings at schools have been nonexistent this last year because of the pandemic, gun violence in general has increased. Philadelphia is already ahead of last year’s record pace in homicides in shootings, and that trend followed nationally in 2020.
“I think it was a brutal combination of people feeling closed out from the world,” Mike Lahiff, president and CEO of ZeroEyes, told Technical.ly. “We’ve seen between COVID, political unrest, the election, the Black Lives Matter movement, and people storming the Capitol last month. There’s a perfect storm of nasty emotions.”
Here’s the breakdown of how the ZeroEyes software works:
- ZeroEyes detects an approaching threat through existing security cameras. Every camera has a real-time streaming protocol data path.
- ZeroEyes’ analytics platform, DeepZero, runs over two parts: pre-processing and inferencing. Pre-processing uses motion detection to capture 10 frames, or 10 opportunities per second to find a gun. Inferencing determines if a gun is present. Every key frame image goes through the ZeroEyes dataset.
- Next, the alert of a weapon goes to the monitoring team or onsite security team.
- If a true weapon is detected, an alert is sent to a local emergency dispatch (such as a 911 call center), onsite security staff, police and school administrators. This process takes three to five seconds and bypasses the traditional dispatch process.
- The software integrates with existing satellite mapping of buildings. As a shooter passes a camera, the map will light up. This allows first responders to know the precise location of a threat. The mapping system can prevent a shooter from entering a building by locking the doors.
The startup was founded and built mainly out of the Pennovation Center in Greys Ferry, but since expanding its team to nearly 40, ZeroEyes moved into new office space in Conshohocken. Lahiff said the company has found its product market fit, and will likely use the new funding — lead by Grateful Investment Group and Legion Capital — to build out teams specializing in the education, commercial and government spaces they’ve expanded into.
And you’ll likely see the headcount rise this year too, he said, with expectations to nearly double employees.
“It is crazy that it takes a pandemic to stop school shootings,” Lahiff said. “We built the system to work with cameras anywhere. A church, a synagogue, local government building, mall — everywhere is low-hanging fruit for someone who wants to go out and create violence. We were fortunate that our system could adapt.”