When video analytics made their debut in the security industry, there was a lot of excitement surrounding their potential and how they could possibly transform the surveillance market. Over time, however, people’s enthusiasm for video analytics waned as many companies overpromised on the capabilities of the technology but failed to deliver a solution that provided consistent results. Of course, the technology itself never went away and various analytic algorithms are used today on a regular basis.
The presence of video analytics companies in the industry though has faded to a degree, as many have either been acquired by or decided to license their patents to hardware manufacturers.There continues to be innovation in the area of analytics, however, and there are a couple of companies looking to make an impression on attendees at this year’s ISC West tradeshow in Las Vegas.
One of these companies is Solink; a Canada-based firm that has developed a contextual analytics solution which helps users combine video data with other enterprise information to provide real-time, actionable intelligence. According to Solink CEO Michael Matta, the company was founded in 2009 and was born out of an ATM skimming issue that investment firm Wesley Clover, which also funds Solink, was having with one its other companies that was working with a large bank in the U.S at the time.
“The company was losing $20 million to $30 million annually from ATM skimming and there was really no way to address it,” explained Matta. “The hypothesis, at the time, was if you can combine video data with transactional data from ATMs, you could help identify when a customer is present and his or her activity at the ATM. That was kind of our first crack at dealing with event-based information.”
Based on the success the company had in the banking sector, Matta said they believed they could take the same principles used in correlating data from an ATM transaction with video and apply it to other vertical markets.
“What we realized is that people could conventionally use video as a time and location-based search. You have information, you want to look it up, but it’s always reactive and you’re always pulling out old data and you have to do the correlation of what camera did this occur at and when did it happen,” Matta said. “We said how can we flip that on its head and say rather than looking at the reactive event of what happened, let’s look at patterns of events. Let’s look at everything that’s happened in the past with similar patterns and then use video as a driver to incorporate the contextual information that’s there, i.e. how long has a person been standing there? How many people are there? What’s the profile of the people that are standing there?”
Matta said the company came up with three tenants pertaining to where their technology could make a difference: turning video into understandable information or event-based information; taking the massive amounts of video information that has traditionally been stored and then deleted at certain point and turning it into an application for the entire enterprise; and, being a software-only company.
Although banking has been Solink’s primary vertical, Matta said that they are focused on what they refer to as the “multi-unit enterprise,” which is essentially any organization that has fewer than 100 cameras per location with dozens to thousands of locations to manage.
“We see the physical security department at a bank as an opportunity and they ultimately own the asset and so we work with lots of physical security managers at banks. But then that opens up another opportunity for us to work with the operations people, the retail banking people or the marketing people at banks because they’ve got this huge asset of surveillance cameras and the applications historically built on top of that architecture are geared towards loss or risk, but we see an opportunity beyond that.”
For example, Matta said a camera above a cash register at a retail store would be used by loss prevention officers to focus on criminal activity, but at the same time, that camera can also be used by store operations personnel to determine the effectiveness of a cashier as well as by the marketing department to see what merchandise is being purchased.
“We think pulling this data back and analyzing it and creating tailored applications by business unit is key,” said Matta. “You need to start with the security person. The security person owns that video asset and they need to see a benefit from the solution before you go off and start selling it to marketing or operations. Today, we only sell a security side and an operations side. We haven’t expanded it to all of these other business units that we could give value to, but we think that’s the opportunity eventually.”
Another company at this year’s show with a unique approach to analytics is Extreme Reality, which is taking its technology to the biometrics market. Based in Israel, Extreme Reality can provide users with a full-body motion analysis of a person through the use of standard surveillance cameras.
“The core technology really relies on our ability to extract the three-dimensional information of a person within a 2D video camera, even the one on your mobile phone,” explained Dor Givon, the company’s founder and chief technology officer. “The way we do it is by knowing it’s a person, looking for a person and overlaying on top of the image we extract the three-dimensional information of the human body. This system now is accurate and robust (enough) to provide solutions that are far more sophisticated than (what’s available) in the consumer electronic world and that’s why, since last year, we’ve started to build and design solutions for the biometric world.”
In the gaming market, for example, Givon said that their software could be used to detect suspicious behavior, such as someone tampering with gaming machines in a casino. “By analyzing the body motion, we can understand if you’re doing something you’re not supposed to be doing,” he said.
Givon said the purpose of analytic solutions like the one his company offers is to consolidate data and turn it into usable information.
“You want to use hours and hours of recorded data in databases and extract information on suspicious people,” Givon explained. “The idea is to extract an amazing amount information from available databases, from exiting cameras and saying, ‘I don’t want to be still, I want to put an additional layer of information working across cameras, across servers, across locations in order to extract data – whether it is in real-time or whether it is in databases.’”
In the future, Givon believes that these types of analytic solutions will become a part of everyday life rather than just a function of security.
“It is going to be a big brother world, with our without us,” said Givon. “It will not only be a security world, it will be a digital identity sort of going with us and we’re being identified by systems around us all the time.”
