Open videobase: A semantic framework for video social networks.

摘要

With the advance of the Web 2.0, video-based social networks such as YouTube have been rapidly gaining popularity. While current video social networks are effective in promoting people's interactions, they are insufficient to represent contents and relationships about videos. It is mainly because of their tag-based static annotations, which oversimplify the meanings of rich video contents with limited semantics of videos.;We propose a new semantic framework, which (1) assists users to annotate videos with a community-based knowledge source, (2) defines annotation graphs to specify topics and types that can represent meanings of videos and their relationships, (3) manages a semantic video social network according to the notion of interest groups. We examined the run-time performance with large-scale simulation data sets and showed analytically that the framework is scalable with respect to the size of the annotation graphs.