| Abstract: |
Collaborative work sources like Wikipedia or
Flickr incorporate semantically organized, non-expert
human knowledge. They appear to be an interesting source of semantic knowledge that could be
exploited in the area of visual recognition. During the work on the project, the student
will try to automatically extract, cross-verify and strengthen the extracted knowledge.
The student will then have an opportunity to use the extracted knowledge in order to
improve the existing visual recognition systems.
|
