Self-Organizing Maps of Web Link Information
We have developed a method that utilizes hypertext link information in image retrieval from the World Wide Web. The basis of the method consists of a set of basic relations that can take place between two images in the Web. Our method uses the SHA-1 message digest algorithm for dimension reduction by random mapping. The Web link features have then been used in creating a Self-Organizing Map of images in the Web. The method has been effectively tested with our PicSOM content-based image retrieval system using a Web image database containing over a million images. The method can as such be used also in other Web applications not related to content-based image retrieval.
KeywordsImage Retrieval Relevance Feedback Image Page Random Base Vector Abstract Semantic Level
Unable to display preview. Download preview PDF.
- 1.Kleinberg, J. (1997) Authoritative sources in a hyperlinked environment, IBM Technical Report RJ 10076, May 1997.Google Scholar
- 2.Brin, S. and Page, L. (1998) The Anatomy of a Large-Scale Hypertextual Web Search Engine, Proceedings of Seventh International World Wide Web Conference, April 1998, Brisbane, Australia.Google Scholar
- 3.FIPS PUB 180-1 Secure Hash Standard (1995) http://www.itl.nist.gov/fipspubs/fip180-l.htm Google Scholar
- 4.Kaski, S. (1998) Dimensionality Reduction by Random Mapping: Fast Similarity Method for Clustering, Proceedings of IEEE International Joint Conference on Neural Networks, May 1998, Anchorage, Alaska.Google Scholar
- 6.Koikkalainen, P. and Oja, E. (1990) Self-organizing hierarchical feature maps, Proceedings of 1990 International Joint Conference on Neural Networks, January 1990, Washington, USA, Vol. 2, pp. 279-284.Google Scholar