Las Vegas – Artificial intelligence (AI), the idea that machines can learn and make decisions on par with a human, may seem like some nebulous concept concocted by Hollywood screenwriters, but the technology itself is already transforming many industries and it is slowly working its way into physical security. In fact, you may have already interacted with AI numerous times and never realized it. For example, if you have asked the personal assistant on your smartphone – be it Siri on the iPhone or “Ok Google” on an Android device – a question or watched a movie or television show on Netflix in the “recommended for you” section, you’ve experienced AI in action. Nearly every tech giant is leveraging deep learning technology to a large extent and most of those organizations are using technology built by NVIDIA.
Long known for its development of graphics processing units (GPUs), NVIDIA has also been one of the pioneers of AI technology. At this year’s ISC West show in Las Vegas, the company is hosting more than a dozen security industry partners in its booth (#20075) highlighting how they are incorporating NVIDIA’s AI platform in various ways to deliver advanced analytics functionality to the traditional safe city and public safety market, as well as the emerging smart city sector, which includes such things as intelligent traffic management applications.
According to Deepu Talla, vice president and general manager of the Tegra business unit at NVIDIA, their goal for the show is two-fold: to educate the broader market – integrators and end-users – about the benefits that AI can provide and show the vendor community how they can incorporate the company’s technology to create these advanced neural networks for security.
“Many industries today are being transformed and, in some cases, disrupted by AI technology,” Talla says. “However, this industry is still not fully aware of the capabilities of deep learning. The leaders within the industry know but the majority of the industry is still not aware of the benefits.”
Given the hundreds of millions of surveillance cameras already deployed globally, which is expected to grow exponentially over the next several years, Talla says humans alone cannot watch these streams and recognize potential security incidents with a high degree of accuracy. Unlike traditional, rules-based video analytics, which Talla says were overpromised and under-delivered, the performance of current AI is almost unmatched, even by human standards.
“Today, most analytics never reach the accuracy of a human [analyzing video] in the same situation,” Talla explains. “But AI has achieved almost super-human accuracy and more progress is being made every day.”
According to Talla, the NVIDIA’s AI platform consists of several products that enable hardware and software vendors to develop security solutions leveraging deep learning capabilities. These include the company’s DGX-1, which Talla says is essentially an AI supercomputer in a box. The system delivers performance equal to 250 conventional servers and comes integrated with hardware, deep learning software and development tools. In addition, NVIDIA’s Jetson TX2, TX1 and TK1 modules can also be embedded in NVRs or even cameras themselves , allowing processing of complex data to be performed at the edge.
The company already boasts an impressive array of industry partners, which range from household names, such as Hanwha Techwin America and Milestone Systems, to relative startups like Robotic Assistance Devices (RAD) which is looking to leverage AI to create an entirely new paradigm within the industry.
“The deployment of advanced analytics continues to play an increasingly important role is the design and deployment of today’s most advanced integrated security systems. Our partnership with NVIDIA aims to escalate the level and utility of analytics to a new level by incorporating artificial intelligence in our video surveillance solutions,” Soon Hong Ahn, president, Hanwha Techwin America, said in a statement announcing their partnership with the company. “We are very pleased to work alongside NVIDIA to co-develop innovative and powerful video surveillance solutions that will deliver the foremost levels of safety and security.”
“By using the NVIDIA Jetson embedded platform, we can develop smarter autonomous robots that can learn from the environment around them to navigate safely, as well as identify and classify objects of interest, including cars and people,” RAD founder and CEO Steve Reinharz said in a statement. “This innovation marks the next generation of robotics.”
While there’s a plethora of AI capabilities waiting to be unlocked for security, Talla says the technology will likely be used initially to improve upon the basic object recognition and tracking analytics with more advanced functionality further down the development pipeline.
“There is a roadmap of things we want to get to,” Talla says. “Prediction, prevention and those kinds of things are the next steps that will happen over the next several years and I think AI is the best solution for tackling those tough problems.”
About the Author:
Joel Griffin is the Editor of SecurityInfoWatch.com and a veteran security journalist. You can reach him at [email protected]
