Proceedings: GI 2010

Seek-n-Tag: a game for labeling and classifying virtual world objects

Bei Yuan, Manjari Sapre, Eelke Folmer

Proceedings of Graphics Interface 2010: Ottawa, Ontario, Canada, 31 May - 2 June 2010, 201-208

  • BibTex

    @inproceedings{Yuan:2010:,
    author = {Yuan, Bei and Sapre, Manjari and Folmer, Eelke},
    title = {Seek-n-Tag: a game for labeling and classifying virtual world objects},
    booktitle = {Proceedings of Graphics Interface 2010},
    series = {GI 2010},
    year = {2010},
    issn = {0713-5424},
    isbn = {978-1-56881-712-5},
    location = {Ottawa, Ontario, Canada},
    pages = {201--208},
    numpages = {8},
    publisher = {Canadian Human-Computer Communications Society},
    address = {Toronto, Ontario, Canada},
    }

Abstract

We identified that virtual worlds that rely on user generated content often lack accurate metadata for their objects. The apparent lack of metadata is a problem for users who are visually impaired, as they rely upon textual descriptions of objects to be present for accessing virtual worlds using assistive technology, such as a screen reader or tactile display. This paper presents a scavenger-hunt game for the virtual world of Second Life -called SEEK-N-TAG- that allows sighted users to label objects as well as collaboratively develop a taxonomy for objects. SEEK-N-TAG aims to build a set of objects with accurate metadata that can be used as training data for an automatic object classifier. Our approach is novel due to its internal approach where the game is implemented in the virtual world itself as to improve its own accessibility. A user study with 10 participants revealed that labeling objects with a game is more effective and accurate than manually naming objects.