Video played a strong role in finding the suspects in the Boston Marathon bombing, but it was mostly old-style detecting that appeared to do the trick as the FBI and police pored through hours of clips collected on bystanders' phones, then put it out for crowd-sourcing help.
Chances are good that the next time an attacker targets a major American city, institutional surveillance will play a more sophisticated role, perhaps in stopping an attack and certainly in sorting out what happened in the event of another. Some of that work, in fact, is already underway.
If last week's bombing woke up the public to the value of widespread street cameras, that wasn't the case for people in the public safety business. On the contrary, even as the runners headed toward Boylston Street and the fateful explosions that rocked the nation, Boston was among a handful of cities moving headlong into a new generation of surveillance technology.
Metro Boston is among the U.S. regions that got multimillion-dollar federal grants in 2012, in part to increase video technology in mass transit and port systems. And Boston is one of the cities moving cameras (and the software that can sort out all that data) from their traditional perch in transit systems into the streets, said David Gerulski, vice president of BRS Labs, a Texas company that installs artificial intelligence systems for video surveillance.
Facial recognition is getting a lot of attention, but that technology still has a ways to go before it can reliably identify grainy images of ballcap-clad, sunglass-wearing suspects. Other forms of analysis include advanced artificial intelligence that actually "learns" what to look for, and software that can boil hours of video down to a few minutes with time stamps to make it clear — made by an Israeli firm with its U.S. headquarters in Farmington.
"Most of these large cities have already been going down the path to do exactly what everybody's wondering if they're going to do," said Gerulski at BRS, whose systems issue warnings when something unusual happens in a monitored scene. "They're not just putting in thousands of cameras, they're putting in tens of thousands of cameras."
Surveillance cameras aren't new, of course. In the years after the 9/11 attacks, cities installed them, civil libertarians fretted about them, comedians joked about them and at least one viral video celebrated the beautiful human moments that they captured — kisses, people helping people, dancing in the streets and even some thwarted crimes.
But when it came to preventing crime and terrorist acts, the networks of cameras did not prove effective. Cities, Gerulski said, "just shut them off."
Why weren't they effective? The problem is Big Data. As anyone who sends video by email knows, moving images create massive amounts of data, and it's all useless unless someone has a way of sorting through it. Humans can do it, but it takes an army of monitors to isolate even one incident. City security forces don't have that kind of manpower.
Learning Like a Child
Now, analytic technology is catching up with our ability to create and store images. And capturing, connecting and storing video is also becoming cheaper by the month, creating, suddenly, a critical mass that's causing a revolution in street-by-street surveillance.
"There is a tremendous field that is opening up and needs to open up regarding surveillance and threat detection to identify terrorists and criminals," said Rik Kirchner, whose Maryland firm, KIRIK International, trains public and private officers to recognize threats.
"It is going to grow exponentially over the next several years," said Kirchner, who until the end of 2012 was the Pentagon's chief of the office of threat detection.
Companies, including Stanley Black & Decker and United Technologies Corp., have in recent years built large and highly profitable security businesses, not only developing technology but integrating other firms' video surveillance capabilities into systems they sell to clients.
Among the more interesting developers is BriefCam, the Israeli firm with its U.S. base in Farmington, whose software summarizes vast amounts of recorded video.
"There was a significant advancement in our own technology, and the entire industry is making great strides in the ability to analyze and handle video," said Amit Gavish, general manager for the Americas at BriefCam.
Gavish would not say whether the firm's technology was used in Boston before the Tsarnaev brothers were identified, or whether it's now being used to piece together the events from available videos. The product is in use by government and public safety agencies in the United States, China, Taiwan, Israel and elsewhere. It was used after the 2011 attacks that left 87 dead in Oslo, Norway, and it's used by such private firms as Stew Leonard's Farm Fresh Foods in Connecticut.
The concept is simple, as Gavish explains it. "If you have 10 hours to investigate on a specific camera, the software will take it to a 10-minute clip … events that occurred during those 10 hours will be presented simultaneously."
Each event is "tagged" and marked with a time stamp on screen, so the viewer is watching events that happened hours apart at the same instant.
Gavish, a former deputy head of security for the office of the president of Israel, compares the technology to Google for the content inside a video. "There is no effective tool to search video ... in that respect, we are the search engine for video," he said.