The most widespread way in which content-based access to video information is supported is through using a combination of video metadata (date, time, format, etc.) and user-generated description (user tags, ratings, reviews, etc.). This has had widespread usage and is the basis for navigation through video archives in systems such as YouTube, Open Video and the Internet Archive. However, there are limitations with this such as vocabulary issues, and authentication across the users who annotate content. Our work has concentrated on addresing content-based video retrieval, the kind of direct retrieval we routinely perform on text, retrieving web pages or blog posts based primarily on actual content.