Wednesday, 29 Apr, 2009 Technology

Study of Flickr Photos Could Help Create Online Travel Books


Researchers from the Cornell University have downloaded and studied about 35 million images posted on the popular website Flickr by over 300,000 photographers from different countries, with the help of a supercomputer installed at the Cornell Center for Advanced Computing (CAC).

Scientists presented their study at the International World Wide Web Conference that took place in Madrid in the period between April 20 and 24. The research offers a new technique of automatically organizing, labeling and summarizing huge collections of digital pictures. The new method's scalability makes it possible to mine data stored in very large collections of pictures, opening ways for creating an online travel guidebook, which would automatically spot the best sites to visit during a vacation, selected by thousands of photographers from around the globe, informs the website of Cornell University.

During their study, scientists also provided statistics regarding the most photographed cities and landmark in the world, collected from the research of the multi-terabyte collection of images.

Here is the list of 25 cities that are the most popular among photographers from Flickr:

1. New York City

2. London

3. San Francisco,

4. Paris,

5. Los Angeles,

6. Chicago,

7. Washington, D.C.,

8. Seattle,

9. Rome,

10. Amsterdam,

11. Boston,

12. Barcelona,

13. San Diego,

14. Berlin,

15. Las Vegas,

16. Florence,

17. Toronto,

18. Milan,

19. Vancouver,

20. Madrid,

21. Venice,

22. Philadelphia,

23. Austin,

24. Dublin

25. Portland.

Here is the list of the most popular landmarks among Flicks users:

1. Eiffel Tower, Paris;

2. Trafalgar Square, London;

3. Tate Modern museum, London;

4. Big Ben, London;

5. Notre Dame, Paris;

6. The Eye, London;

7. Empire State Building, New York City.

Cornell researchers came up with methods of identifying the most photographed locations, listing results for thousands of places at city as well as landmark scales.

"We developed classification methods for characterizing these locations from visual, textual and temporal features. These methods reveal that both visual and temporal features improve the ability to estimate the location of a photo compared to using just textual tags," said Daniel Huttenlocher, the John P. and Rilla Neafsey Professor of Computing, Information Science and Business and Stephen H. Weiss fellow.

According to CAC director David Lifka, the acceleration of digital information will make supercomputers along with high-performance storage systems important instruments for quickly storing, archiving, keeping and recovering large-scale collections of information. It is worth mentioning that the study was supported by the National Science Foundation (NSF) and funded by Google, Yahoo! and the John D. and Catherine T. MacArthur Foundation. The Cornell Center for Advanced Computing is supported by Cornell, the NSF, the Department of Defense, the Department of Agriculture and affiliates of its corporate program.

Powered by

Add your comment:

antispam code