An indexing strategy without recording temporal relationships for video database systems

In video databases, each video contains temporal and spatial relationships between content objects. Moreover, the information related to locations and motions of objects is considered in video databases as well. One of the well-known video indexing strategies is the 3D C-string strategy. However, it cannot index for the condition that an object appears and then disappears for more than one time. Moreover, since the 3D C-string records the relative locations of objects, it cannot effectively derive the absolute locations of objects. Therefore, in this paper, we propose 2D Video String indexing strategy. The strategy can resolve these two problems happened in the 3D C-String strategy. Moreover, we only build index for spatial relationships, and we can still answer all types of video queries, i.e., object, spatial, temporal, and spatio-temporal queries. From our simulation study, we show that our proposed strategy is more efficient for video search than the 3D C-string strategy. A prototype video information system is also implemented to show the effectiveness of indexing video data based on our proposed strategy.