Agent Vi’s patented, all-software platform efficiently partitions the analytics workload between the edge and the network. The camera processes the original captured video and transmits results onto the network as ultralow bandwidth data packets. Agent Vi’s virtual management software, which can run anywhere on the network, processes Agent output packets and generates meaningful events based on rules defined by the user.
The Agent Vi software contains a module that performs the raw processing of images as seen by each camera in the field prior to any image compression. This process is called "feature extraction." The result is very low-bandwidth data (usually less than 20Kbps) which is transmitted to the server for further processing.
The "brain", and main component of the software, sits anywhere on the network and receives and processes data streams from any number of cameras with the embedded agent connected to the network. Using the patented distributed processing architecture, Agent Vi performs a wide range of software pre-configured detection missions in real time.
All the system data, including event detection rules and detailed event history, resides on a central database that also provides backup, retrieval, and extensive reporting capabilities.
Agent Vi is designed for easy integration inside video devices such as IP cameras, DVRs and video encoders.