Indexing video information with user-generated bookmarking

Gaskin, Nathaniel (2009) Indexing video information with user-generated bookmarking. BSc dissertation, University of Portsmouth.

[img] PDF
Restricted to Registered users only

Download (3416kB)

    Abstract

    Textual information on the web is readily searchable, with search engine spiders able to extract salient semantic details using both the text content and the mark-up of the page to build indexes of content. However, as bandwidth and storage become ever cheaper, video is quickly gaining momentum as a primary medium through which to share knowledge over the Internet. The problem lies in the fact that this video-based information cannot be spidered in a similar way to text, and is therefore not searchable. Whilst several machine learning techniques exist that attempt to discover the information or semantic concepts present within video, they face many challenges that limit their widespread use. This project proposes a means through which videos are added to a system, and then 'bookmarked' and 'tagged' with searchable keywords at interesting points by the users of the system. By assuming these users to have more domain knowledge about the contents of the video than any automatic transcription mechanism, we can rely on them to build indexes of the important facets of the videos within the system. In doing this, we allow these videos to become searchable by other users of the system based solely on user-generated content and description.

    Item Type: Dissertation
    Departments/Research Groups: Faculty of Technology > School of Computing
    Depositing User: Jane Polwin
    Date Deposited: 20 Jan 2011 12:48
    Last Modified: 28 Jan 2015 11:15
    URI: http://eprints.port.ac.uk/id/eprint/671

    Actions (login required)

    View Item

    Document Downloads

    More statistics for this item...