This is an unusual moment in time to be writing about a homeland security-related investment. War in the Middle East has gripped headlines for the past week, temporarily pushing homeland security off the front page – but not for long. The issues attendant to homeland security, such as border security and perimeter defense, continue to be of paramount importance, and relevant technological solutions present major investment opportunities, though not without major risk.

For the canny investor, the challenge is to source out markets not yet disintermediated, locate first generation technologies on the verge of their next generation and gain first-mover advantage.  Not by being the first to market, but by being the first significant player in that market with a technology that really works. In the case of analog video surveillance and video analytics, we’re witnessing a classic case of a market evolving right before our very eyes — from the legacy system’s analog, closed circuit television (CCTV) technology to the costly, error prone first generation of video analytics to today’s all digital next generation enterprise-grade video analytics.  And it’s the last category – the next generation of enterprise-grade video analytics that has recently attracted attention from the end use customer community.  Not only has 21Ventures invested in digital video analytics with Aspectus Video Intelligence, Inc. — Red Herring reports that at least eight other startups raised funds this year, including: 3VR Security, EnVysion, VideoNext, and Vigilos.
 
Video surveillance is a market that has yet to be disintermediated in 30 or 40 years, largely employing analog, CCTV technology provided by legacy vendors despite the post 9/11 public’s heightened awareness to security issues. So, with today’s interest in increased security within both the public and private sectors and the advent of video analytics, why has this market been stymied for almost four decades?

Let’s go back in time and look briefly at the evolution of video surveillance.
From the early 60s thru the early 90’s, video surveillance consisted primarily of CCTV being watched by human operators on multiple monitors, while simultaneously being recorded on analog tapes for historical backup and forensic purposes.  Considered revolutionary in its day, two main problems quickly became apparent with CCTV video surveillance.  First, academic and empirical studies proved repeatedly that humans can’t watch multiple TV monitors with comprehension for more than 20 minutes per hour.  And second, the footage was forensic in nature. CCTV provided hours of taped records, but featured no capabilities to analyze them in real-time and trigger event alarms. 

In the early 90’s, “Computer Vision” evolved to meet these two problems.  Originally a venue for research labs, the government and academia, “Computer Vision,” as video analytics was first called, evolved out of artificial intelligence research.  The initial commercial offerings by first generation video analytics venders were based on video motion detection (VMD) algorithms developed by the U.S. Defense Advanced Research Projects Agency (DARPA).  All of these early attempts were costly, required large amounts of hardware and bandwidth and, unfortunately, delivered extremely high false alarm rates.

And then came 9/11, which obviously caught everyone by surprise, including the VC community who was just beginning to sense the need for video analytics in the defense and enterprise markets.   Sobering urgencies of the post-9/11 era delivered swift government and some VC investment in first generation video analytics that could detect abandoned objects, human intrusion of restricted areas, stopped cars, etc. — all benefits not provided by legacy CCTV video surveillance systems.

Like their earlier VMD cousins, today’s first generation video analytics architecture requires costly hardware, cameras each with servers attached to one or multiple cameras in the field, plus the fact that they consume prohibitively large amounts of bandwidth.  Added to this, customers and venders find all first generation video analytic solutions to be complex and error-prone with frequent performance problems, specifically, low probability of detection (PD), high false alarm rates (FAR) and system crashes. 

Another major problem also emerged.  Who within the enterprise and government organizations would manage and support these new systems — traditional physical security departments, or the IT network managers?   This problem continues to this day. When video analytic venders call upon end use customers, it is still unclear who the ultimate decision maker is.

Unfortunately due to these problems, today’s video analytics marketplace remains frustrated, while demand continues to soar.  A few startups eagerly posture in the wake of these first generation video analytics vendors, but there is no established leader in a marketplace that remains anemically penetrated as the entire industry anxiously awaits a low cost, low bandwidth and error free standard.

Case in point.   According to Frost & Sullivan, The market is exploding as both the public and private sectors, including the enterprise, now understand the potential value of video analytics.  There are an estimated 80-100 million surveillance cameras worldwide, most of them “dumb” cameras in need of video intelligence technology.   Frost and Sullivan estimates that the video surveillance software market will expand at a compound annual growth rate (CAGR) of 25% over the next six years and will reach $670.7 million annually by 2011.   UK-based IMS Research states that the world market for video analytics software is expected to increase from its current level of $67.7 million to $839.2 million in 2009, at a 65.5% CAGR.

What’s needed by the security industry is a completely new approach. Our belief at 21 Ventures is that, given the worldwide need to secure the public and private sector from the threat of terrorism, the time has come for a next generation enterprise-grade video analytics solution that overcomes all of the above mentioned problems.  One that’s cost effective, but more to the point, one that’s field-tested for reliability, low bandwidth enabled, interoperable and completely scalable.  Most importantly, this digital architecture already exists and has been successfully used worldwide since 2003.  In the case of our portfolio company Aspectus, it’s an innovative image processing architecture from the worlds of IT and artificial intelligence that resolves the problems of low PD and high FAR, at a reasonable cost and without bandwidth overload.

The issue of who handles enterprise security is also undergoing a paradigm shift. As protocol solutions move from analog CCTV to digital technology and video analytics gets employed by enterprise networks, 21Ventures believes that ownership of the video security network will be assumed by the IT departments.  This is a current roadblock.  Physical security departments continue to resist because staffers don’t necessarily understand (Internet Protocol) IP technology, nor do they want to give up control of surveillance to the IT department.  But given that security is a multi-trillion dollar global business, of which hundreds of billions are spent annually on security personnel, being more effective and efficient is a tremendous incentive.

Many enterprises and defense departments have security infrastructures that are located all over the world.  How do you manage that in an effective and efficient way?  If the alarms are carried over an IP network, the organizations have the ability to centralize the service, reduce security personnel and greatly improve response time.  This technology already exists with the next generation video analytics and has been field tested and proven for the past three years.  For example, Aspectus has successfully deployed a multi-station IP-based video analytics solution for Israel Railways.

According to Frost & Sullivan, “Network-based systems are expected to completely replace analog and hybrid (analog video capture devices with digital storage devices) systems by 2015-2016, providing huge opportunities for video management software.”

As investors know, it’s hard to find IT industries that haven’t been disintermediated, and this one hasn’t.  For video analytics, a new all digital approach to video intelligence architecture will jump-start the next generation. And in four or five years time when the industry is completely converted from analog to digital, the next-generation technology leader will be the one to have set the industry standard.