The neural network learning algorithm presented in the paper splits the problem of handwritten digit recognition into easy steps by learning character classes incrementally: at each step, the neurons most relevant to the considered class are fixed so that subsequent classes will not disrupt the knowledge already acquired, but will be able to use it. A new relevance measure is also defined, for which a cheap approximation can be computed. The advantage of the attribution scheme starts to show even for small experiments, but should become more obvious as the number of classes increases. Picking only a few relevant features for each class, and sharing them between classes, constrains learning and improves generalization. Experiments were limited to pre-segmented digits, but our use of a spatio-temporal network architecture makes their extension to unsegmented strings straightforward.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Incremental character recognition with feature attribution


    Contributors:
    Audouin, R. (author) / Shastri, L. (author)


    Publication date :

    1995-01-01


    Size :

    488488 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Incremental Character Recognition with Feature Attribution

    Audouin, R. / Shastri, L. | British Library Conference Proceedings | 1995


    Number Plate Recognition Based on Multi Feature and Character Recognition

    Thanojkumar, S. / Suneetha, I. | BASE | 2017

    Free access

    A Feature-Based Neural Wavelet Optical Character Recognition System

    Iftekharuddin, K. M. / Jemili, K. / Karim, M. A. et al. | British Library Conference Proceedings | 1995


    A feature-based neural wavelet optical character recognition system

    Iftekharuddin, K.M. / Jemili, K. / Karim, M.A. | IEEE | 1995


    Recognition and attribution of variable message signs and lanes

    Nienhuser, Dennis / Gumpp, Thomas / Zollner, J. Marius et al. | IEEE | 2008