Athena Security's Network application flow diagram for active shooter detection system:

We use a protocol that is embedded in every camera. This video stream feeds from any type of camera into the edge compute or mobile device where it runs the model. The edge or mobile device are capable of running the stage 1 and stage two algorithms, model loosely after the human brain, that are designed to recognise patterns. They interpret sensory data through a kind of machine perception, labeling or clustering raw input.  After the edge device finds a match it will send it to the 3rd  network for final verification thought the cloud.  The application backend then sends an alert tailored to your configuration settings either to mobile app, VMS, or 3rd party monitoring systems.