Dragonfruit AI, a provider of AI-powered video analytics, on Thursday officially unveiled Dragonfruit Frontier, a new cloud video management software offering featuring embedded analytic capabilities. Frontier allows end-users to leverage their existing video surveillance hardware infrastructure while simultaneously enjoying the benefits of a state-of-the-art VMS.
For just under a base price of $1,000 per site per year, Frontier was designed to provide an affordable, full-featured VMS offering to multi-location enterprises. That price includes the company’s Base Station hardware, licenses for up to 20 cameras, local or cloud video storage, advanced features, such as video walls, real-time alerts, and access control integrations, as well as all the available analytic capabilities on-demand.
“We have really focused our VMS to be built for multi-location deployments. We are helping our customers securely manage hundreds of locations from a single browser. The idea is that this out of the box works for multi-location, there is no additional charge for [analytics], and if you want to scale globally, you can,” Amit Kumar, the Founder and CEO of Dragonfruit AI, explains. “As opposed to someone who is focused on, let’s say one location with 1,000 cameras – on-premises video analytics companies, for example, do a lot of those kind of deployments – that is not our sweet spot. We think that the cloud and the ability for you to do analytics in the cloud and be able to do it in a way where you can solve bandwidth problems and so on, means that it is possible to take video intelligence and make it scale.”
Whereas many analytic solutions in the past were thought of as purely a high-end feature on an expensive surveillance project, Kumar, a Silicon Valley veteran who previously founded and sold two other software companies, says they can provide the same capabilities at a much more lower price point, thereby democratizing the technology for a wider array of users.
“The key here is the completeness of the product. If you look at the products we have, they are way beyond, most of the [competitive set],” he adds. “And not only that, it includes all of the video analytics as part of the package."
Simplified Deployment
The company’s Base Station is built on the Apple M1 platform, a new ARM-based system-on-chip (SoC) from the tech giant that promises to revolutionize the GPU market by providing increased video processing power at a much lower price.
“It is really, really inexpensive, but it is really, really powerful,” Kumar says. “We are in a partnership with them, so the Base Station basically gets shipped directly from Apple to every location.”
The Frontier software comes embedded on the Base Station, so once it is opened at the site, all the installer has to do is plug it in to a power and Ethernet connection and configure it remotely from a browser.
“There is no other configuration that is required on-premises. That makes the deployment super easy,” Kumar explains. “If you’re going to scale something to hundreds and thousands of locations, that only works if the work that you need to do per location is low and so we’ve made it really simple.”
Traditional and Custom Analytics
The robustness of the analytics they offer as well as their ability to develop custom algorithms is another thing that differentiates Dragonfruit AI, says Kumar.
The company’s suite of analytics covers a wide range of traditional offerings, such as people/vehicle detection and license plate recognition, as well as a variety of newer applications, including face mask detection, slip-and-fall recognition, and various algorithms that can help organizations with compliance issues vis-à-vis liquid spill detection for retailers and hard hat detection for construction sites. Kumar says they are also highly regarded for their bespoke analytic creation, which are eventually made available to all their customers.
“The way our pipeline works is that things that we do bespoke for a customer, in a quarter or two become available for all of our customers,” Kumar explains. “Liquid spill detection is a good example of something that started off as a bespoke model and now it is available to any customer who is interested in using it. When we started the company a long time ago, the vision was always to build one software stack, one software platform that covers all the needs that a physical security installation might have.”
The key to Dragonfruit AI’s ability to provide this suite of video analytic offerings lies in its patented Split AI technology, which enables customers to scale analytics while keeping bandwidth requirements low.
“You see a lot of [proof of concepts] that never lead to actual sales for video analytics companies on large-scale environments because… people end up doing things in the cloud and they develop a [proof of concept] that works well on one or two cameras, but as soon as you go up to 10 cameras, 20 cameras or 100 cameras, you don’t have enough bandwidth. The problem is there is always a push-and-pull with IT where they don’t want security to interfere with operations,” Kumar says. “We can solve that problem with Split AI, which reduces the bandwidth costs dramatically. This technology that we have built takes AI processing and splits it between the Base Station and the cloud. We shift all of that GPU hardware to the cloud but keep a little of our processing on-prem.
“What does that look like? Let’s say you deploy a spill detection module. If there are no spills going on, then all of this video is flowing to the Base Station and the Base Station is observing that video, so it is not sending anything to the cloud,” he continues. “But let’s say there is a spill, then what happens is that when this video gets to the Base Station, the Base Station does a basic analysis of the scene and says, ‘I think there is a possible spill here, but I don’t have enough processing power and Dragonfruit doesn’t want me to have huge processing power so that I would cost a lot, so what I’m going to do is send some of this information to the cloud.’ Then all of this information or some of this information gets sent to the cloud and the cloud says, ‘aha, there is a spill, let me dispatch an alert.’”
According to Kumar, although the company speaks with end-users frequently, they exclusively go to market through the integrator channel.
“Generally speaking, we work with integrators and distributors around the world, so we have partners not only in the U.S., but in Canada, South Africa, Australia, India, Thailand, Singapore, so we have a very good network of partners,” he says.
To learn more about the company, visit https://www.dragonfruit.ai/.
Joel Griffin is the Editor-in-Chief of SecurityInfoWatch.com and a veteran security journalist. You can reach him at [email protected].
