One of the biggest trends in video surveillance over the past several years has been the use of deep learning solutions to derive enhanced intelligence from video data. Often referred to artificial intelligence or AI-powered analytics, today’s video analysis tools are finally starting to deliver on the promises of some of the technology’s forerunners that were present in the industry nearly a decade ago.
But while many people think these analytic solutions are simply plug-and-play, the reality in most cases is much different, in fact; setting up some of today’s AI-based tools can require a significant amount of time and skill, something that keeps them out of reach for much of the market. However, Lumeo, a San Francisco-based startup, is looking to change this by offering the industry a “no code” analytic builder that enables users to setup their own custom analytic algorithms without the need for advanced technical skills.
“We’re the first if not the only video analytics platform that empowers businesses and solutions providers to create and deliver completely custom video analytics in a matter of minutes and with a predictable, affordable price in mind,” explains Devarshi Shah, the company’s Co-Founder and CEO. “We’re trying to lift the entire industry up, enable companies and solution providers to leverage the billion-plus cameras that exist out there today… and bring analytics from this complex thing that only a few people are able to do, to now something that everybody’s able to leverage and have solutions for.”
Chris Bruce, the other Co-Founder and CTO of Lumeo, which recently received the SIA New Product Showcase “Best New Product” Award, said that they came up with the idea for Lumeo after seeing the challenges that exist for organizations across a wide range of industries when it comes to incorporating video technology into their products. For example, Bruce wanted to add a camera to a wearable baby monitor product at a previous company where he worked that was eventually sold to toy giant Mattel, but he realized there was no easy way to do it.
“Just to add a camera to our product, there was no off-the-shelf solution for it and for a company like Mattel, which is primarily a toy company and not necessarily a technology company, it become this overwhelming thing trying to build it in house,” Bruce says. “That was one of the main points where we set out to say, ‘hey, this needs to exist’ and our algorithms predicted when babies would wake up from a nap and told parents when top put them down and things like that. We wanted to use a camera to do those analytics visually and, again, it was very difficult.
“After that, Devarshi actually had a hardware company at the time as well and we both had the same shared investor and we started talking about this idea of doing a smart neighborhood watch system,” he continues. “What we wanted to do was use everybody’s cameras in a neighborhood, whether it was Ring or what have you, and collectively work together and use AI to look for anomalies, to look for people that didn’t belong, and patterns that didn’t belong. Again, we ran into the same problem: there is no platform just to get connected to the camera, offload this data and pipe it through machine learning models and things like that. It just didn’t exist… so we decided to build it ourselves.”
How Lumeo Works
According to Shah, one of the key differentiators for Lumeo is that they allow customers to build any kind of analytic.
“We are a full stack analytics platform. In other words, we interface with cameras, NVRs, VMS systems, and we allow these analytics to run and we curate a wide range of regularly used AI models, so you can pretty much build any type of solution that you can imagine. Think of them as building blocks that you can put together for the analytic,” Shah says.
Secondly, Shah says Lumeo enables users to do all of this with “drag and drop” tools so they can customize how they want particular analytics to work in a user-friendly way. “We enable relatively non-technical users to be able to pull these analytics together very easily and, at the same time, the platform is fully expandable using APIs, custom code, and so if you have some engineering staff in house, you can build your own analytics completely custom right within the platform.”
Shah also emphasized that Lumeo’s analytics can run on the edge or in the cloud.
The Commoditization of Analytics?
Obviously, the Covid-19 pandemic has given rise to a variety of new security products that have been launched into the marketplace over the past 18 months, including things like face mask detection which were not readily available when the coronavirus began spreading across the country in early 2020. With Lumeo, however; Bruce says users could have developed their own analytic for this and other Covid protocols and cut out the middleman.
“The difference is that Lumeo provides a platform where non-technical folks, the people that are actually in operations, could build that themselves without having to understand AI, how to stream video around or any of that stuff. They can just build a use case for it themselves,” he adds. “We are the first ones that have really built a tool that allows system integrators, solution providers or even end-users to solve their own problem. We have all the building blocks, so it is just like using Legos to build a custom analytic and today, the market is all about vertical analytics. We’ve now just turned analytics more into a commodity where you can build a custom one and, through our application, we actually allow people to create solutions and publish them so that if you already have a solution you built that other people can use, they can just use that without having to create something from scratch.”
Lumeo’s pricing, which starts at $10 per camera per month, also promises to bring analytics to the wider surveillance market.
To learn more, you can find Lumeo at booth #10056 during ISC West 2021 this week in Las Vegas.
About the Author:
Joel Griffin is the Editor of SecurityInfoWatch.com and a veteran security journalist. You can reach him at [email protected].
