Umbo offers groundbreaking AI-based video security solutions
Umbo Computer Vision is an artificial intelligence startup that develops autonomous video security products specifically for enterprises.
The firm’s award-winning product, Umbo Light, is an autonomous AI-based video security system that can be accessed from both desktops and mobile phones. The end-to-end device uses rapid neural networks and the most advanced AI technology available to provide organizations with qualified, accurate alerts about security risks in real-time to keep people safe, prevent crimes, deter intruders and protect valuable assets stored in both private and open areas.
Effectively, Umbo Light is a smart camera that works without physical on-premise recorders. The device stays connected vis-a-vis network connections and through the cloud. Multiple cameras can be networked and managed from multiple locations through a single Web login, which has been used more than 1.5 million times by customers.
The firm has designed the device to relay data to a cloud-based, neural net based AI system in order to protect people and physical assets through interactive learning. As the device observes the environment, it is designed to provide smarter alerts over time.
The firm, which builds both the camera’s hardware and develops and maintains the firmware and cloud system’s algorithms, is led by a former DARPA robotic challenge U.S. champion and a da Vinci robotic surgical system researcher, and recently raised US$9.6 million in two venture capital rounds.
The company’s CEO believes the product’s big advantage is that its leverages proper machine learning capabilities, which results in accurate data and a lower number of false alarms, especially when dealing with complex scenes and weather conditions.
He notes that Umbo Light’s AI-based cloud technology, real-time autonomous video capabilities and learning cameras are designed to provide enterprise-class customers with end-to-end protection, as well as fast, accurate alerts.
“When we set out to develop Umbo Light three years ago, our vision was to provide our customers with an autonomous video solution that enabled a safer and better world,” said Shawn Guan, CEO of Umbo Computer Vision. “With these new advancements, we are delivering on our promise to reimagine how video can understand what it sees in real-time the same way that humans do. This enables our customers to protect the people and things that they care most about — at all times.”
The technology is deployed by a network of 200 active system integrators and has over 500 active end-users using it across 13 physical security verticals. Umbo’s AI-based video security system has built its intelligence by analyzing over two billion security images. According to the firm’s own research, 12 cameras or more per operator is too much visual data for a human to process. Based on this approximation, 400 million cameras would require 120 million operators to monitor them. Umbo’s algorithm allows the monitoring task to be automated and conducted in the cloud instead, accessible online and through API integration.
The firm’s video-based security system can determine occupancy, loitering, undertake a person tally and distinguish people in mass event scenarios, including football tailgating parties.
The firm bills its “AI events” system as a simple way to apply highly complex AI models to security systems at an enterprise scale. “The key differentiator for Light,” says Guan, “is that Light actually understands human behavior, not just lines or boxes on a screen. Unless someone is actually scaling a wall, you won’t hear about it.”
In addition to tracking motion, Umbo Light learns about behaviors from what it observes so it better understands a user’s needs the more it is used at a specific location. The system uses AI to sequence images and also to compare the relevance of the images to one another. The result is a a temporal end-to-end solution that can alert users about critical security events or provide real-time data that can allow for even preventive action.
Due to the technology’s uniqueness, the firm won a Best in Video Analytics Award from the Security Industry Association (SIA) in 2017 and also won an a startup of the year award from NVIDIA, a computer graphics processor manufacturer.
The company’s products are currently in use at schools, government facilities, and many commercial sites to protect critical infrastructure and safeguard the public. The product is also used in small offices, local stores, hotel franchises, large national retailers, warehouses and at major transportation facilities.