Collection of Material Related to Content-Based Image Retrieval (CBIR)
- Limitations of Content-based Image
Retrieval This is the latest (final?) version, posted on
June 11, 2008. (The first draft was posted in June 2007 and the second
in May 2008.)
- Web Test Results of Content-based
Image Retrieval. (Appendix A)
- Simulating What is Happening
in Current CBIR systems (Appendix B)
- Color Histograms Provide
No Semantic Information (Appendix C)
- Edge Histograms Provide No
Semantic Information Either (Appendix D)
- "The Holy Grail of Multimedia
Information Retrieval: So Close or Yet So Far Away?" - An Answer
A three page summary of the issues discussed in the form of an answer
to a question raised in an editorial in the IEEE Proceedings.
- ICPR08 Presentation
Slides and Notes for the presentation "Limitations of CBIR"
in the 2008 International Conference on Pattern Recognition, Tampa,
Florida, Dec. 8-11.
- A Set of Images for Testing CBIR
Techniques and a Collection of Results Posted December
20. Replaces earlier postings about a "Challenge to CBIR".
- An Evaluation of the Scale Invariant
Feature Transform (SIFT) Draft posted August 25, 2008.
- The Number of All Possible Meaningful
or Discernible Pictures Draft updated May 7, 2009.
- A Hidden Obstacle to Image Retrieval
Draft posted May 22, 2009. Latest revision June 5, 2009.
- "Why Meaningful Automatic Tagging of Images is Very Hard" (Proc. ICME2009,
pp, 1432-1435) Slides
and Zipped Slides
(July 1, 2009).
External Links
Image-Net A
database of images organized according to the WordNet hierarchy, in which
each node of the hierarchy is depicted by hundreds and thousands of images.
You can enter a word and obtain all images in the database labeled by
that word. The results provide a powerful demostration of the semantic
abyss. In my view the best database for CBIR research. Paper
describing the system. (Link entered on May 8, 2009.)
Index modified on July 1, 2009.
Back to List of Technical Writings
|