GunDetectTM Helping Security Cameras become Intelligent Video Threat Detectors
GunDetect is part of an intelligent video threat detection system that can be deployed in areas where security is a concern, accuracy is required and authorities need to be alerted to potential threats as fast as possible. Security Teams visually scanning multiple camera feeds for criminal, violent intent is time consuming, costly and frequently results in missed events. With a system designed to complement and assist human review by initially filtering of all your video feeds and highlighting when guns are unholstered, your Security Team can concentrate on those video feeds where specific attention is needed. Further human review can then be focused on assessing intent and handling the situation.
So how does it work? We hook into your existing camera systems and analyze the video feeds in real-time. If a gun is detected, we notify your Security Team that further review is required. The result is typically several times faster than detection by humans who are scanning multiple video camera screens. And we are very cost effective, under $100 per camera per month.
NanoWatt Design: Understanding the Visual World with Computers
NanoWatt Design, the parent company of GunDetect™ is changing the way companies are interacting with video cameras by reviewing the visual data in each video frame for specific patterns. This visual data is then extracted and our computers are trained to find specific objects from the video frames.
We do this with our Visual Computing Platform (VCP) that processes, analyzes and understands the frame images from a video camera in real time. It can then automate a response based on our computer‘s vision perception of that specific image.
Plus it is state-of-the-art accurate. Our Visual Computing Platform (VCP) is able to positively detect a pre-defined object (for example: guns) in an image that is in motion within the camera's field-of-view for a minimum of 2 seconds, 90+% of the time. And using machine learning technology, the more images it looks at, the more accurate VCP gets. If the pre-defined object is within the camera view for 10 seconds or more, pre-defined objects will be missed less than 5% of the time.
GunDetect was started as NanoWatt Design wanted a real-world example of how visual object detection can work on our Visual Computing Platform (VCP). It is a perfect example of how computers and security cameras can complement human vision to provide additional information on a situation, resulting in a better outcome for everyone.