A Novel Method for Mining Video Association Rules Using Weighted Temporal Tree

With the ever-growing digital libraries and video databases, it is increasingly important to understand and mine the knowledge from video database automatically. Video association mining is a relatively new and emerging research trend used to discover and describe interesting patterns in video. The traditional classical association rule mining algorithms can not apply directly to the video. It differs in two ways such as, spatial and temporal properties of the video and significance of the video sequence items. Most of the video association rules mining algorithms discover frequent item sets considers only temporal properties that do not consider the quantity in which items have been appeared in the video sequence. This paper discusses an efficient method for discovering a weighted temporal association rules from a large volumes of video sequence data in a single scan of the database using Weighted Temporal Tree structure. Video association rule consists two key phases are (i) Video pre-processing and (ii) Video association rule mining. The preprocessing phase converts the original video sequence into a temporal video transaction format. The mining phase consists three tasks namely, weighted temporal tree construction, frequent pattern extraction and rule mining. The proposed weighted temporal tree based association rule mining did not require multiple scans. The mined association rules have more practical significance and identifies the valuable rules comparing with Weighted Tree based algorithm. We also presented results of applying these algorithms to a synthetic data set, which show the effectiveness of our algorithm.