Content-based semantic associative video model

Existing methods for representing video are not capable of expressing relationships among the objects in shots. However, semantic networks are advantageous to expressing semantics of videos and improving the efficiency of video retrieval. This paper proposes a semantic associative model based on semantic networks, it represents the contents of a video shot by semantic objects and their relationships that appear in the shot. By representing the semantic relationships among objects which appear in different shots, this model can express the semantic correlations among shots. Using the semantic relationships, we can retrieve videos easily.