Francisco Carrero, José Carlos Cortizo, Borja Monsalve (Wipley, Social Gaming Platform)
José Carlos Cortizo Pérez <josecarlos.cortizowipley.com>
Sometimes, when you’re searching through Flickr, it’s better to search in multiple languages in order to get more results. Since Flickr is a site whose popularity has spread throughout the world, millions of users upload picture titles and tags in their native tongue. This site will allow you to insert a search query for Flickr, and it will search for you in multiple languages. This greatly broadens the scope of results you’ll get. For instance, search for the word “car”, and the site will search for it in many different languages, allowing you to see pictures of a car someone form Turkey might have uploaded, which you might otherwise not have been able to find. A simpler way to look at it is as a mash-up of Flickr search and Google’s translation services.
FlickrBabel is a images search engine that retrieves images from Flickr that matches the user’s query. FlickrBabel first expands the query by adding the translation of the query to 5 languages (English, Spanish, German, French and Portuguese) and then sends the expanded query to Flickr and shows the images retrieved. FlickrBabel also shows a map for the geolocalized images and allows the users to restrict the results to a given location (city, area, country, etc.).
FlickrBabel is developed in PHP and makes use of Google Translation and Flickr APIs.
The system’s needs are a Web Server with PHP support, constant Internet access and a developer’s key to Flickr API.
There are 4 main modules: 1) query pre-processing and expansion, 2) search process, 3) visualization and 4) geolocation restrictions.
There are some R&D projects related to multilingual multimedia retrieval but there are no user-friendly tools that implements the techniques related to those projects. FlickrBabel is a simple but user oriented solution.
FlickrBabel is the first product of Social Gaming Platform, first created as a experiment to test Google Translation and Flickr APIs, but finally implemented as a mainstream Internet application.
- There are no scientific references that directly explains FlickrBabel, but part of the work behind FlickrBabel has been applied in:
- Carrero, F., Cortizo Pérez, J. C., Gómez Hidalgo, J. M., Testing Concept Indexing in Crosslingual Medical Text Classification. Proceedings of the IEEE ICDIM 2008.