Architecture & Engineering Consultant Liaison Program
The BRS Labs "A&E Consultant Liaison Program" provides tools and support to assist consultants who conduct risk assessments and security audits, and design and specify security systems.
The AISight Value Proposition
AISight is the only reason-based behavior recognition product on the market, and is protected by over 60 patents and/or patents pending. Since this totally unique application can scale without limits, it is the only realistic solution for detecting abnormal behavior at large deployments that have hundreds or thousands of camera feeds.
AISight Architecture
AISight software runs on standard platforms and integrates with standard storage, compression, and database technologies. The technology can integrate with communications switches and customer relationship management software, as applicable, from multiple vendors across both traditional and next-generation communications networks. The system can perform completely self-contained; however, it is usually beneficial to have a network connection to provide remote support.
AISight integrates with existing security surveillance systems and only requires a video stream. Video streams are read by AISight using the RTSP protocol for reading digital streaming video in MPEG-4 and H.264 formats. Other video formats can be used, but a poor quality video input will result in reduced quality analysis of the scene. For customers who do not already stream their video with these standards, numerous hardware solutions from various vendors are available to convert analog or digital video streams into the necessary digital format.
AISight's deployment architecture uses a two-part observe-and-learn cycle to build observation models, learn patterns, and perform its analysis. The observe portion of the architecture, the VAE, observes events, records a time stamp, and classifies event participants (humans, dogs, cars, etc.). The learn portion of the architecture, the MLE, takes input from the VAE and builds up a series of Hypocepts™ models , which are abstract models of observed behavior. These Hypocepts™ models are then used to compare current observations from the VAE with established, learned patterns. This comparison results in an abnormality measurement known as a "temperature" of how unusual the current activity is compared to the learned behaviors.
As video data is transmitted by the security cameras, the data is passed along to the AISight Analysis Server, which interprets the data and determines if the area of view is normal or abnormal. Following this analysis, if AISight identifies an abnormality, an alert is generated on the AISight system manager application server. At this point, the server analyzes the alert and decides where to send it. Alerts can be sent to a variety of end points, including a system manager console, an AISight operator console, PDAs, and iPhones.