Finding visual concepts by web image mining

ABSTRACT

We propose measuring "visualness" of concepts with images on the Web, that is, what extent concepts have visual characteristics. This is a new application of "Web image mining". To know which concept has visually discriminative power is important for image recognition, since not all concepts are related to visual contents. Mining image data on the Web with our method enables it. Our method performs probabilistic region selection for images and computes an entropy measure which represents "visualness" of concepts. In the experiments, we collected about forty thousand images from the Web for 150 concepts. We examined which concepts are suitable for annotation of image contents.

REFERENCES

• 1: Kobus Barnard , Pinar Duygulu , David Forsyth , Nando de Freitas , David M. Blei , Michael I. Jordan, Matching words and pictures, The Journal of Machine Learning Research, 3, 3/1/2003
• 2: Keiji Yanai, Generic image classification using visual knowledge on the web, Proceedings of the eleventh ACM international conference on Multimedia, November 02-08, 2003, Berkeley, CA, USA [doi>10.1145/957013.957047]
• 3: K. Yanai. Web image mining toward generic image recognition. In Proc. of the Twelfth International World Wide Web Conference, 2003.
• 4: Keiji Yanai , Kobus Barnard, Image region entropy: a measure of "visualness" of web images associated with one concept, Proceedings of the 13th annual ACM international conference on Multimedia, November 06-11, 2005, Hilton, Singapore [doi>10.1145/1101149.1101241]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION

General Terms:
Algorithms, Experimentation, Measurement

Keywords:
image recognition, probabilistic method, web image mining