M. Deepika, K. Dhanalaxmi, B. Karthika and I. Persi Pamela
The number of videos on the Internet is growing explosively, which poses a great challenge on automatic video search. Existing video search engines, such as YouTube, visualize the search results in a ranked format. The simple list structure may be suitable for locating a movie trailer or a music video, but cannot handle queries with complex topic structures so it creates difficulties for the users to locate their needed information. In this case, a semantically structured hierarchy with nodes covering all the critical topic facets becomes a much better choice. Obviously, manually creating topic hierarchies for every query is impractical simply due to the great number of the search topics. We use a Prototype-based Hierarchical Clustering algorithm to create an automatically generate hierarchical structure by processing the video descriptions along with their semantic information. We extract this information from a semantic database using a sub-ontology concept because we can’t have the whole semantic database in our website.
Semantic Database, Sub-Ontology, Prototype- Based Hierarchical Clustering