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.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Applying the information bottleneck principle to unsupervised clustering of discrete and continuous image representations


    Beteiligte:
    Gordon, (Autor:in) / Greenspan, (Autor:in) / Goldberger, (Autor:in)


    Erscheinungsdatum :

    01.01.2003


    Format / Umfang :

    535257 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Applying the Information Bottleneck Principle to Unsupervised Clustering of Discrete and Continuous Image Representations

    Gordon, S. / Greenspan, H. / Goldberger, J. et al. | British Library Conference Proceedings | 2003


    Image Segmentation by Unsupervised Sparse Clustering

    Jeon, Byoung-Ki / Jung, Yun-Beom / Hong, Ki-Sang | IEEE | 2005


    Convolutional Patch Representations for Image Retrieval: An Unsupervised Approach

    Paulin, M. / Mairal, J. / Douze, M. et al. | British Library Online Contents | 2017


    Information bottleneck through variational glasses

    Voloshynovskiy, Slava / Kondah, Mouad / Rezaeifar, Shideh et al. | ArXiv | 2019

    Freier Zugriff

    Bottleneck

    Forschungszentrum Jülich / University of Wuppertal / Cologne University | DataCite | 2009