Anomalous stationary object detection and reporting

This invention describes a computer system that watches a video feed to identify objects that remain still for an unusually long time. It learns what typically happens in a scene, like cars moving or people walking, and then calculates a 'rareness score' for anything that stays put. If an object's stillness is rare for that location and duration, the system flags it, and it gets smarter over time by observing what usually happens.