Semantic Relatedness for Knowledge Transfer
License
The provided software is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Germany License. The data provided has no specific license as it was created from Wikipedia, Flickr, or the WWW, i.e. sources with diverse licenses/copyright restrictions.
Citation
The accompanying publications:
- www.d2.mpi-inf.mpg.de/rohrbach10cvpr [1]
- www.d2.mpi-inf.mpg.de/rohrbach10eccv [2] (Parts and Attribute Workshop, Code will be available soon, send me an e-mail if you would need it soon :)
Talk
Software & Data
The software is provided as Java Eclipse Project and uses JUnit (included in standard Eclipse distribution) and was tested under Linux and Windows but should run on all systems where Java is available.
Please refer to the README in the software and data package for further details.
Requirements
Eclipse
with JUnit (should work without both, but makes life much easier)
WordNet
Obtain jwsl.jar from grid.deis.unical.it/similarity/
More specifically: e-mail JWSL as subject to Giuseppe Pirrò: gpirro [at] deis [dot] unical [dot] it
Yahoo
Yahoo BOSS API key from developer.yahoo.com/search/boss/
Flickr
Flickr API key from www.flickr.com/services/api/misc.api_keys.html
Wikipedia
Download the ESA index for Wikipedia from below
Download
- Software and data package (version 0.1) (2.8 MB)
- ESA index for Wikipedia (4.6 GB)
Help & Contact
If you need any help or have any suggestions feel free to contact me.
References
[1] What Helps Where – And Why? Semantic Relatedness for Knowledge Transfer, M. Rohrbach, M. Stark, G. Szarvas, I. Gurevych and B. Schiele, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June, (2010)
[2] Combining Language Sources and Robust Semantic Relatedness for Attribute-Based Knowledge Transfer, M. Rohrbach, M. Stark, G. Szarvas and B. Schiele, Trends and Topics in Computer Vision: ECCV 2010 Workshops, October, Volume LNCS.6553, p.15-28, (2010)