Social computing, folksonomies, and image tagging: Reports from the research front

Abstract

AbstractToday, the World Wide Web contains a vast collection of multimedia information sources which are mainly created and published by individuals. Image and video sharing services such as Flickr and YouTube pose new challenges in multimedia information indexing and retrieval and demand dynamic sets of solutions. Individuals are not just users but creators and indexers of multimedia information sources through activities such as tagging. User generated tags, taxonomies, and folksonomies of images are gaining popularity. Therefore, how to incorporate users' tags into the process of indexing and retrieval needs urgent attention. Presenters on this panel conducted various studies on image tagging and descriptions by people and the common findings include: (a) tags assigned to groups of images and individual images differ in terms of their level of abstraction, (b) image tagging specificity and exhaustivity levels differ greatly among individuals, and (c) the accordance between existing controlled vocabularies and tags varies in terms of image attributes.To make this panel an interactive one, the audience will have the opportunity to discuss the general topic of social computing (specifically image tagging) and indexing guided by the following questions: Are user‐generated image tags/index terms viable alternatives to index terms assigned by professionals? Or, if tags cannot substitute for indexer‐generated index terms, can we utilize user‐generated tags to complement current indexing schemes by reflecting users' needs which are not found in existing indexing schemes? Can a large image indexing and retrieval system realistically rely on user‐generated tags/index terms? What does the audience think of the criticism that language used by users (or user‐generated tags/index terms) is not precise and controlled? Are we in a better position today, due to the popularity of social tagging, to effectively and efficiently implement truly user‐centered indexing? Or, is social tagging merely a fad that users will not support in the long term? What should the role of traditional indexing tools and systems be?