Detecting anomalous events using a long-term memory in a video analysis system

摘要: Techniques are described for detecting anomalous events using a long-term memory in video analysis system. The may be used to store and retrieve information learned while system observes stream of frames depicting given scene. Further, the configured detect occurrence events, relative observations other that have occurred scene over time. A distance measure determine between an active percept (encoding observed event depicted frames) retrieved previously memory). If exceeds specified threshold, publish review by users