Security cameras have become valuable tools to identify threats of violence. There's a new development Wednesday that adds artificial intelligence that can detect guns in security video.
The new dimension of artificial intelligence, or AI, has the potential to prevent a mass casualty event.
"We needed to move away from the thinking that emergency mass notification systems were all about notification, " said David Fraser, CEO of Omnilert.
Fraser demonstrated its new Gun Detect technology it claims to be an industry first using AI. Utilizing existing camera networks in a large-scale setting, such as a school or airport or large business campus, algorithms can identify a gun or shotgun. The system first flags the gun in orange as suspicious, then goes to red alert in seconds.
"They basically can make a high-resolution determination of whether this is a real threat or not," said Fraser.
In the demonstration for ABC7 News, a suspect with a face mask walks into view of a security camera, briefly casing out the room and then pulling out a gun. Artificial intelligence identifies it as a threat.
"After that," noted Fraser, "it's a single click to be able to initiate a predefined set of actions."
A cell phone pop-up notifies a security officer. Software then provides a still photo to get details of the gun and a description of the suspect. Then a decision can be made whether to dispatch first responders or to dismiss it as a false alarm. AI knows, for example, that a sprinkler head is not a gun.
Low-light conditions or an impaired field of vision could impact AI's ability to recognize a threat. However, AI is showing promise of early warning before shots are fired.
Mike Leone is a senior analyst at Enterprise Strategy Group.
"Over time and as we continue to collect video and as we better train some of these algorithms and models and to best detect those images or objects, I think it's only going to get better," he said.
Omnilert sees college campuses as a key market for its AI technology.