Domain Specific Video Data Mining Using Multicore Fp-Growth Tree

摘要

Content based methods are necessary when text annotations are non-existent or incomplete. Furthermore, content based methods can potentially improve retrieval accuracy even when text annotations are present by giving additional insight into the media collections. For domain specific videos, associations among different events are used for indexing purpose, which contributed a lot in bringing vast knowledge from video libraries and databases. In this paper, we are introducing multicore Fp (frequent pattern)-growth tree building algorithm applicable to streamed data. This algorithm will improve efficiency to retrieve videos through parallelizing frequent pattern constructing steps.