BRS Labs says it’s no longer competing with video analytics software
|BRS Labs' AISight|
After winning the “Best Video Analytics” award at the ISC West security conference for two consecutive years, BRS Labs now declares that its “behavioral analytics” software should not even be compared to video analytics, which it believes has failed to win a place for itself in the security market.
“We don’t compete with video analytics anymore,” said John Frazzini, president of BRS Labs, at a press conference on Oct. 12 at the ASIS security conference in Dallas, TX. “We’re not interested in having that conversation.”
BRS Labs has developed artificial intelligence software, dubbed AISight, which takes visual inputs from a video surveillance system and learns to distinguish “normal” behavior at a particular location from “abnormal” behavior. As BRS Labs described it in a press release it issued last March, its software “automatically learns and stores what activities and behaviors are normal for each view, and generates real-time alerts that notify appropriate security personnel of potentially threatening behaviors, all without human involvement or training.”
For example, imagine a video camera – which was using the AISight software from BRS Labs -- was aimed at a parking lot in a suburban office park for three months and took images of thousands of passenger cars entering and exiting that parking lot, but never saw a single truck enter or leave the lot. The software would “learn” that it was normal for cars to enter and exit that parking lot, but highly “abnormal” for a truck to do so. If, someday, a truck was to drive into that parking lot, the BRS software might trigger because it had concluded that a suspicious “abnormal” event had suddenly occurred.
This approach to spotting threats is different from traditional video analytics, which relies on the operator to set pre-determined rules for acceptable behavior, and will trigger if someone or something breaks those pre-determined rules. For example, an airport authority might decide that it was appropriate for pedestrians to walk down a particular ramp and out a specific “exit” door, but that it would be unacceptable for the same pedestrians to walk in through that exit door and up the same ramp. Those rules could be established when the video system was first installed, and the analytic system would trigger whenever the camera saw someone violating those rules.
Frazzini said the security industry was now recognizing the value of BRS Labs’ approach to behavioral analytics. “We’re being embraced by the industry,” he added, noting proudly that Boeing and SAIC had recently tested his company’s software and that both giant systems integrators had begun to include BRS Labs’ software in some of the video surveillance proposals they had submitted to their potential government customers.
In fact, Frazzini said two contenders for a huge surveillance contract with the U.S. Army Corps of Engineers – SAIC and Johnson Controls – had included BRS behavioral analytic software as part of their bids. The outcome of that procurement has not yet been announced, said Frazzini, but he believes that SAIC will be named the winner.
He also announced that another company, VidSys, was integrating AISight into its physical security information system (PSIM), including its situation management application, known as RiskShield.
BRS Labs also boasted about a recent new contract with “one of the largest businesses in Latin America,” which Frazzini declined to name, which would involve the company’s behavioral analytics software being employed by approximately 15,000 video surveillance cameras at about 3,000 different locations in South America.
Frazzini was remarkably candid in saying that many video analytics systems fail to live up to their hype. That’s the “big elephant in the room that no one discusses,” he added. “Not a lot of value is being delivered by the industry.”
Frazzini has quite a different view of the future of artificial intelligence when connected to video surveillance systems. He said the BRS Labs approach has been used by U.S. defense agencies, intelligence agencies and nuclear facilities, and that the reaction from end-users has been “overwhelming.”
Frazzini and his colleagues at BRS Labs apparently feel their time has come. “I believe these technologies will be ubiquitous in five or 10 years,” Frazzini concluded.