Applying the information bottleneck principle to unsupervised clustering of discrete and continuous image representations
- 1 January 2003
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 370-377 vol.1
- https://doi.org/10.1109/iccv.2003.1238368
Abstract
We present a method for unsupervised clustering of image databases. The method is based on a recently introduced information-theoretic principle, the information bottleneck (IB) principle. Image archives are clustered such that the mutual information between the clusters and the image content is maximally preserved. The IB principle is applied to both discrete and continuous image representations, using discrete image histograms and probabilistic continuous image modeling based on mixture of Gaussian densities, respectively. Experimental results demonstrate the performance of the proposed method for image clustering on a large image database. Several clustering algorithms derived from the IB principle are explored and compared.Keywords
This publication has 8 references indexed in Scilit:
- Clustering artPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Feature representations for image retrieval: beyond the color histogramPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Blobworld: image segmentation using expectation-maximization and its application to image queryingPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Unsupervised document classification using sequential information maximizationPublished by Association for Computing Machinery (ACM) ,2002
- A Continuous Probabilistic Framework for Image MatchingComputer Vision and Image Understanding, 2001
- Elements of Information TheoryPublished by Wiley ,2001
- Hierarchical browsing and search of large image databasesIEEE Transactions on Image Processing, 2000
- Color indexingInternational Journal of Computer Vision, 1991