A Novel Semantic Tagging Technique Exploiting Wikipedia-based Associated Words

Abstract

With the rise in popularity of smart devices such as smartphones and tablet PCs, the volume of multimedia contents is growing exponentially through social networks and multimedia contents sharing systems. However, the characteristics of mobile devices and tags lead to the performance degradation of the tag-based image retrieval. In this paper, we propose a novel semantic tagging method exploiting Wikipedia-based associated words. The proposed system consists of two phases. In association words construction phase, extraction manager parses the page information from Wikipedia dump, and then semantic analyzer constructs the associated words using disambiguation page. In tag recommendation phase, a user selects single meaning among various associated words, so that associated words which belong to the chosen cluster are automatically assigned to the contents. The proposed system makes it possible to annotate the multimedia contents with the reliable and relevant tags in the smart device environment.