Action parsing using context features

Nagita Mehrseresht

We propose an action parsing algorithm to parse a video sequence containing an unknown number of actions into its action segments. We argue that context information, particularly t…