A view of the BRS Labs booth at ISC West 2013. BRS Labs President John Frazzini believes much of the analytics being offered by hardware vendors is a "gimmick" intended to sell more cameras.
Photo credit: (Photo courtesy Joel Griffin)
The concept of video analytics can stir up a wide range of reactions among both manufacturers and end-users in the security industry. Many people still roll their eyes at the thought of the technology, which had once generated a lot of enthusiasm about the potential it had to revolutionize video surveillance. The problem, of course, is that many companies overpromised the capabilities of analytics, but ultimately failed to deliver a product worthy of the hype that it generated. There are a plethora of reasons why the technology failed to ultimately take off the way a lot of people in the industry envisioned it would and a number of companies that used to have a substantial presence on the tradeshow floor at ISC West are no longer around.
That’s not to say, however, that analytics have completely evaporated from the video surveillance market. Quite the contrary, many in the industry have once again become receptive to the idea of using some type of analytic functionality in their organization – be it people counting, heat mapping, object left behind, or virtual tripwires – as a way to improve safety or business operations.
According to Axis Communications General Manager Fredrik Nilsson, humans are clearly better at performing intuitive tasks than cameras are which is what got analytics providers in trouble nearly five years ago with their inability to deliver the capabilities that people expected.
“No one wants a camera, they want a safer environment,” Nilsson told members of the media at a press event on Wednesday.
This is one of the reasons that Nilsson believes end-users were “eager to jump on the bandwagon” of video analytics back then and why they were ultimately disappointed when vendors failed to deliver. Nilsson said that three things really needed to take place before there could be widespread adoption of video analytics; better image quality, enhanced processing power and good algorithms, all of which have now come to fruition.
“It’s not hard to predict what will happen, but when it will happen,” Nilsson said of technology trends. For example, Nilsson pointed out that Apple Founder Steve Jobs had the idea of everyone possessing a smartphone about 10 years before the devices gained widespread adoption.
One company that has been hard at work developing better algorithms for video analytics is German-based camera maker Mobotix. While Mobotix has offered people counting and heat mapping capabilities for some time, Steve Gorski, the company’s general manager for the Americas, said that they will soon be releasing MxAnalytics Activity Sensor, which has the ability to decipher actual movement within a scene from false alarms generated by rain, wind and other natural occurrences.
“We think that, for certain applications, this is going to be a game changer,” he said.
This functionality will be available on new Mobotix cameras without any additional fees. Gorski believes that the Activity Sensor will be ideal for facilities in the oil and gas sector, which have cameras that operate under extreme environmental conditions.
Herve Fages, senior vice president for Schneider Electric’s video line of business buildings business, said that the need for video analytics has always been there, but that the technology came out “prematurely” and was oversold at the time. Fages added that people are giving the technology a second chance and that it has gotten to the point where it makes sense under certain circumstances.
“At the end of the day, the (video) industry is going to go to more analytics,” said Fages, who believes that analytics will eventually become a differentiator that sets companies apart from one another. Another reason that Fages believes there will be increasing developments around video analytics is that organizations simply don’t have the manpower or the budgets to support continuous monitoring of their video feeds.
Anthony DeStefano, director of Schneider Electric’s integrated security sales buildings business, said that it’s virtually impossible to have people monitor hundreds of video streams and that “analytics help you sort it out.” And although the technology may have been oversold when it first arrived in the market, DeStefano said that analytics have improved since that time.
Fages said that the market may have also turned the corner in terms of customer education as people no longer have unrealistic expectations.
“It was vaporware,” Dr. Edward Wassal, vice president of global marketing for DVTel, said of video analytics beginnings in the industry. “All of these things are now real.”
Three years ago, Israeli-based DVTel acquired ioimage, which was one of the forerunners of video analytics technology, to bolster its end-to-end offerings in IP video.
Wassal said that he would eventually like to push some form of analytics capability to the edge with the company’s Quasar camera line.
“Analytics, when it first came out had big promise and it failed,” said Shahar Ze’evi, senior product manager at Tyco Security Products.
Ze’evi said he views analytics as more of a “help tool” for users to convert data into information rather than the radical, life-changing innovation that some people made it out to be years ago. He does believe, however, that video analytics can be successful at an application level as long as people specifically define the problem they’re looking to solve with the technology.
“If you are true to what you have and put an application around it, it will work,” he said. For example, Ze’evi explained that using facial recognition in a small, low traffic application would work out much better than trying to use it in wide open, crowded spaces like airports.
An advancement in surveillance technology that people credit with helping bring credibility to analytics once again is improved image quality. Rick Ramsay, senior product manager for Avigilon, which is a leader in the high-definition video space, said that he too believes that analytics are “slowly” coming back to prominence.
Ramsay added that HD and megapixel surveillance products enable users to find things in footage that may have been missing before, which only helps to improve analytic capabilities.
“I think it’s starting to work well again,” Dr. Bob Banerjee, senior director of training and development for NICE System’s security division, said of video analytics.
Rather than focusing on doing a lot of different things with analytics, Banerjee said that people have realized they just need to do one thing very well. For most users, Banerjee said that analytics boils down to the basic premise of alerting them to the presence of a person at a location where they shouldn’t be.
“Solve the problem by focusing on one thing and doing it brilliantly,” he added.
John Frazzini, president of BRS Labs, which offers video analytics software that has the ability to learn what is normal or abnormal activity within a field-of-view versus traditional rules-based analytic algorithms, has a completely different view of the analytics market.
He believes that video analytics as they previously existed are not really on the rebound and what is largely being offered today is a “gimmick” on the part of hardware manufacturers to sell more cameras.
“There is a collaborative, disingenuous approach to this in the industry,” he said.
Unlike his company’s software platform that sells to enterprise-class users for millions of dollars, Frazzini said he doesn’t see people really buying surveillance hardware products for their analytics capabilities. In fact, he compared some vendors to “modern-day snake oil salesmen.”
Though he admitted that BRS Labs caters to a higher-end customer and that there could be a greater demand in other markets for analytics such as people counting, Frazzini said what’s currently being offered is really more of an add-on feature and not a true revenue generator.