Face Detection-Person Detection Tie-In

When face detection and person detection are enabled at the same time, face objects will be associated with the appropriate person objects. This allows SAFR to continue tracking people even if they turn their faces away from the camera. In addition, face recognition metadata will be used to automatically enhance the metadata of the associated person object. Each face object can be associated with at most one person object. Similarly, each person object can be associated with at most one face object.

When a face object and person object have become associated with each other, events that are generated by the person object are called “parent events” or “root events”, while events generated by the face object are called “children events” or “secondary events”. You can choose if secondary events are included in the Event Archive by enabling or disabling the Include Secondary Events setting on the Events Preferences tab in the Desktop Client. SAFR’s default behavior is to not include secondary events, as they can create too much useless “noise” in the Event Archive.

Face detection and person detection can be enabled and configured on the Detection Preferences tab on Windows machines.

Shared Event Attributes

When a face object and a person object become associated, they will share the following event attributes:

Whenever a face event is updated with any of the above properties, the associated person event will be updated as well.

Secondary events (i.e. associated face events) have their rootEventId attribute set to the eventId of their parent event. (i.e. the person event it’s associated with) This enables all secondary events to be gathered and appropriately presented. Each secondary event has only one rootEventId.

Conversely, root events (i.e. associated person events) have their hasSubEvents attribute set to true.

Root events aren’t ended until all their child events are ended.

See Also