The paper describes a method for automatically extracting informative feature hierarchies for object classification, and shows the advantage of the features constructed hierarchically over previous methods. The extraction process proceeds in a top-down manner: informative top-level fragments are extracted first, and by a repeated application of the same feature extraction process the classification fragments are broken down successively into their own optimal components. The hierarchical decomposition terminates with atomic features that cannot be usefully decomposed into simpler features. The entire hierarchy, the different features and sub-features, and their optimal parameters, are learned during a training phase using training examples. Experimental comparisons show that these feature hierarchies are significantly more informative and better for classification compared with similar nonhierarchical features as well as previous methods for using feature hierarchies.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Feature hierarchies for object classification


    Contributors:
    Epshtein, B. (author) / Uliman, S. (author)


    Publication date :

    2005-01-01


    Size :

    565682 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Feature Hierarchies for Object Classification

    Epshtein, B. / Ullman, S. / IEEE | British Library Conference Proceedings | 2005


    Statistical Learning of Visual Feature Hierarchies

    Scalzo, F. / Piater, J.H. | IEEE | 2005


    Classification hierarchies: a proposed graphical interface

    Rackley, S. L. / Waugh, L. / Froese, T. M. et al. | British Library Conference Proceedings | 1998


    A Feature Level Fusion Approach for Object Classification

    Wender, Stefan / Dietmayer, Klaus C. J. | IEEE | 2007


    Light field distortion feature for transparent object classification

    Xu, Y. / Maeno, K. / Nagahara, H. et al. | British Library Online Contents | 2015