Cashing in on Searchable Video and Analytics

Jan. 20, 2006
Intelligent DVR company 3VR lands $10 million in venture capital - what does that mean for you?

On Tuesday, 3VR Security (, a San Francisco-based video surveillance firm, announced that it had raised a pile of cash -- $10 million to be exact.

The funding comes largely from big-time venture capital firm Kleiner Perkins Caufield & Byers (which brought in Colin Powell as a "strategic limited partner" back in July 2005), but also includes investments from VantagePoint Venture Partners and CIA investment arm In-Q-Tel.

Why the investment? KPCB partner Ted Schlein is calling the 3VR technology, which parses and searches video data and organizes video by person and location, "the most innovative technology coming out of Silicon Valley today." Or as is simpler to explain: Venture capitalists chase after and explore what are predicted to be high-growth areas.

So, what kind of problem does the company's technology purport to solve? The concept is essentially one of time management: If the number of cameras used at today's businesses and by police and government continue to rise, we could soon face a major problem in that we're recording too much video to be able to sort through that video and find the relevant data. We eventually get lost in the piles of videotapes or the hours of MPEG footage if we have enough cameras recording.

Steve Russell, the company's president and co-founder, explains that the 3VR solution is a search-and-sort technology that parses through the video and organizes it by camera, by subject (it uses biometric recognition technology to group clips of what the tools sees as the same person), and alerts with real-time events and create "rules" in the recorded area. It picks up on video analysis functions similar to what the industry has seen from companies like Cernium and ObjectVideo.

Or, as 3VR's co-founder and Executive Vice President Tim Ross explains, the 3VR appliance and its internal technology serves the same function as a web search engine.

"The easiest way to think about 3VR search is to take the Google analogy," says Ross. "Google pre-indexes all 8 billion web pages into an organized set of metadata. When your Google 'San Francisco Giant's home schedule', Google doesn't go out to 8 billion pages, instead it works against their index. 3VR similarly pre-indexes all video within the system looking for motion, direction, objects, people, faces etc. When you search for 'Tim Ross', 3VR is fast searching this index. When you query 3VR for 'show me anytime you think you've seen Tim Ross over any bank branch during the past month', we can search those results in seconds even though there might have been 300,000 hours of video during that time/camera span." caught up with 3VR's president and co-founder Steve Russell by phone this week to talk about what the investment means and where 3VR is headed today.

According to Russell, a "good chunk" of the $10 million in Series B funding will be focused on research and development (the rest will be spent on a sales push). There is a continual challenge of surfing the new technology wave, says Russell, that any cutting-edge technology company faces, especially in a developing area of technology like 3VRs, which merges video storage, video searching, biometrics, and networking into one package.

Not that there hasn't been plenty of R&D already. Having launched the company in 2002, the 3VR product, he explains, has had to see a number of improvements since its original model. He explains that developers have focused on making the product turn-key in nature to facilitate installations. He says they've also wrestled with issues of scale and how much video can be processed, and are now in a distributed type of design where the technology isn't focused on creating analytics at one camera, but instead processes video from any camera across the network. The other challenge, he explains, that this kind of technology has faced has been one of the "analytic pipeline." "People used to buy one very powerful server and have it do one thing," says Russell. "Now the technology searches and delivers multiple uses across a variety of video feeds."

But even with great claims for scale improvements, pipeline improvements, turn-key installations and even the establishment of what the company calls an "affordable" price point, venture capitalists won't invest if they don't think security directors, dealers and system integrators are going to "bite" on the company's technology.

Russell explains that the investment from In-Q-Tel, which was announced in late 2005, in some ways can serve as a stamp of approval for new technology like his company's products. And he adds that 3VR wasn't in a rush to get out of the gate. Instead, he says the company focused on creating a project that wasn't "labware", one that could be as easily deployed to the commercial sector as the government marketplace.

"We think the industry has made a problem for itself by focusing too much on getting technology out the door and getting it to the government," says Russell. "We made a conscious decision early on to focus on creating a viable commercial product."

Getting ROI on Video Analytics

SIW asked Russell to explain how he understands the return-on-investment (ROI) model behind video and search analytics. His summary is a great page for your PowerPoint, whether you're a security systems integrator explaining video analytics ROI to a security director or a security director explaining the ROI of the equipment to your bosses at the company's C level.

ROI #1: Real-time monitoring. "One person can monitor about 10 video feeds for only about 45 minutes before losing concentration," says Russell. Beyond that, they need a break. If they're dealing with more than 10 feeds at a time, the likelihood is that they could be missing something. To deal with a multi-camera installation where you're viewing video from dozens of cameras, you start to face issues of staffing and the costs of hiring alert persons to watch cameras. With today's systems like the one their company is monitoring, the goal is that you can cut that back to one person, using alarm type operations (crossing an imaginary line, an object moving, etc.)

ROI #2: Savings on investigations. Being able to archive and search through volumes of video cuts down on investigations time.

ROI #3: Stopping crime before it happens. By using video analytics to free up security personnel and by receiving immediate reports for non-normal activity, your staff can stop the crimes before the perpetrator is successful. By stopping crimes on time and immediately, you're showing the best kind of ROI there is - protecting people and property before something happens.