Semantic Extraction of Basketball Game Video Combining Domain Knowledge and In-Depth Features

Figure 1

Schematic representation of basketball video actions (the spatiotemporal aggregation of a set of action primitives on the appearance and motion flow of the video in our method. For example, basketball shoot can be represented as a collection of appearance features corresponding to “player,” “ball,” and “basketball hoop” and motion features corresponding to “run,” “jump,” and “shoot”).