Base Security Enhanced by Video Analytics

Irvine Sensors Corporation State of the Art Video Analytic commercial product (ALERT®), which is based on exploitation of cognitive-inspired processing techniques, enables real-time alerting of potentially threatening actions by detecting, tracking, and classifying objects and objects activities in areas that require surveillance to ensure critical infrastructure protection and safety. Spatial, temporal, and color correlations of observed objects and object activities are combined with deep learning techniques to achieve industry best probabilities of threatening detection/classification with very low false alerting rates. The GPU-based processor which is optimized for Air Force security needs can simultaneously process up to 100 video channels in a 1U electronics processor. The project will design this class of processor during phase 1 and will demonstrate it under realistic operating conditions during Phase 2.