Community-Contributed Media Collections: Knowledge at Our Fingertips - Community-Built Databases: Research and Development - page 44

of user-contributed media, enhancing the user searching activity on media collec-
tion, and automatically classifying Wikipedia articles.
However, less attention has been paid to the integration of heterogeneous data
available on different online social networking sites to tailor personalized multime-
dia services. For example, users can explore and gain a broader understanding of a
context by integrating the huge amount of data stored in Wikipedia, YouTube, and
Flickr. Digital libraries which cover a specific field (e.g., geography, economy) can
be enriched by extracting related information from Wikipedia. Similarly, news
videos can be integrated and contextualized with information provided by Wikipedia
articles. Only few approaches have been proposed to address these issues [ 54 , 55 ].
The problem of integrating Wikipedia and a geographic digital library is pre-
sented in [ 54 ]. The integration approach consists of identifying relevant articles
correlated to geographical entries in the digital library. The identification is carried
out by analyzing List_of_ Wikipedia pages, where G is a geographical entity

(e.g., region, country, city). Finally, additional information is extracted by parsing
infoboxes content of selected Wikipedia pages. Experimental evaluation on the
extraction of relevant articles and metadata information show good performance in
precision and recall.
The integration of news videos and Wikipedia articles about news events is
addressed in [ 55 ]. The method aims to automatically label news videos with
Wikipedia entries in order to provide more detailed explanations about the context
of the video content. Using “Wikinews,” a sister project of Wikipedia, all news-
related articles are extracted from Wikipedia, while information about videos is
extracted from the content caption (CC), which is usually provided. The CCs of news
stories are labeled with Wikipedia entries by evaluating date information. Experi-
mental results show that Japanese news videos broadcast over a year were accurately
labeled with Wikipedia entries with a precision of 86% and a recall of 79%.
A host of technical challenges remain for better exploiting the community-
contributed media into personalized applications able to tailor multimedia services
according to the context in which the user is currently involved. Multimedia
services for mobile devices could be personalized by exploiting both the current
context of the user and the huge amount of media content available on the online
social networking sites. For example, WikEye [ 56 ], a system for mobile technol-
ogy, is able to retrieve interesting information on touristic places from Wikipedia
spatial and temporal data. Future research directions might also exploit relevant
knowledge available in different media collections.

G

< >

References

1. Youtube website. http://www.youtube.com/
2. Flickr website. http://www.flickr.com/
3. Wikipedia website. http://www.wikipedia.com/
4. Alexa the web information company. http://www.alexa.com/
5. Paolillo, J.: Structure and network in the youtube core. In: hicss, p. 156 (2008)