New Security Industry Association (SIA) member Nirovision provides powerful artificial intelligence (AI) features like facial recognition and object detection; the company‘s mobile-first video analytics provide real-time AI insights for professionals on the go. Nirovision is headquartered in Sydney, Australia, with a presence in Las Vegas, Nevada, and New York City.
SIA spoke with Jimmy Lee, CEO of Nirovision, about the company, the security industry and working with SIA.
SIA: Tell us the story of your company.
Jimmy Lee: A few years ago, a small group of engineers were working together at a B2C company that sold cloud-connected wireless security cameras to homes and small businesses. One of the most challenging problems that the company faced was dealing with false alarms caused by the motion detection algorithms used by the cameras. Motion detection wasn’t smart enough to distinguish between activities of interest versus mundane things like moving shadows, light changes or the rustling of leaves.
This caused a problem that was twofold: customers were bombarded with false alarm notifications and engineers had to build complex and costly infrastructure to store alarm footage that no one was watching. Coincidentally at that time, deep learning was starting to become a feasible technology thanks to the proliferation of cost-effective GPUs.
With that, Nirovision was founded with the goal of using deep learning to solve vision-based problems encountered by security cameras.
What solutions/services does your business offer in the security industry? And what makes your offerings/company unique?
JL: Nirovision provides deep learning-based video analytics for the security industry. We offer both solutions for the end user and solutions for software developers to build using our AI software development kits and APIs.
Generic AI recognition tools are often trained using high-quality photos from professional photography such as photos of scenery, people posing or product shots. These images don’t accurately represent security camera footage and fail to produce useful results.
Nirovision’s computer vision models are trained using security camera footage to get real-world accuracy. Security cameras are typically mounted in high places and need to track subjects moving across a scene. By focusing on what matters most, such as people and vehicles, our models aren’t distracted by random objects that do not matter.
What’s something we might not know about your company – or something new you’re doing in security?
JL: We’re taking a “mobile-first” approach in developing our end-user product, Niro Aware. I think this is unusual in the security industry, as most security software is a Windows desktop experience with legacy code bases that dates back to Windows 95.
By taking a mobile-first approach, we want to rethink the whole interaction of the end-user with video footage. Slick and intuitive mobile experiences should not be limited to consumer apps alone, so we want to freshen up the security industry this way.
Business today is also increasingly becoming more mobile. We’re seeing fewer and fewer security setups where a guard sits behind a desk to monitor cameras. The power of AI is in notifying you immediately when something happens, so it makes perfect sense to get pinged on your mobile while you’re on the go instead of sitting at your desk waiting for things to happen.
SIA: What does SIA offer that is most important to you/your company? And what do you most hope to get out of your membership?
JL: With us being headquartered in Australia, SIA membership gives us excellent visibility and connections to the U.S. security industry. We look forward to connecting with others in the industry, keeping on top of industry news and attending events.
SIA: How does your organization engage with SIA? What are your plans for involvement in the next year?
JL: We’ll be exhibiting at ISC West 2019. We’ll be at booth 1098, which is coincidentally adjacent to one of SIA’s booths! We’re also participating in the SIA New Product Showcase, so wish us luck!