Subspace methods such as PCA, LDA, ICA have become a standard tool to perform visual learning and recognition. In this paper we propose representational oriented component analysis (ROCA), an extension of OCA, to perform face recognition when just one sample per training class is available. Several novelties are introduced in order to improve generalization and efficiency: (1) combining several OCA classifiers based on different image representations of the unique training sample is shown to greatly improve the recognition performance. (2) To improve generalization and to account for small misregistration effect, a learned subspace is added to constrain the OCA solution, (3) a stable/efficient generalized eigenvector algorithm that solves the small size sample problem and avoids overfitting. Preliminary experiments in the FRGC Ver 1.0 dataset show that ROCA outperforms existing linear techniques (PCA, OCA) and some commercial systems.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Representational oriented component analysis (ROCA) for face recognition with one sample image per training class


    Beteiligte:
    De la Torre, F. (Autor:in) / Gross, R. (Autor:in) / Baker, S. (Autor:in) / Vijaya Kumar, B.V.K. (Autor:in)


    Erscheinungsdatum :

    01.01.2005


    Format / Umfang :

    1215689 byte





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Component-based LDA method for face recognition with one training sample

    Jian Huang, / Yuen, P.C. / Wen-Sheng Chen, et al. | IEEE | 2003


    Component-based LDA Method for Face Recognition with One Training Sample

    Huang, J. / Yuen, P. / Chen, W.-S. et al. | British Library Conference Proceedings | 2003


    Managing the Roca electrification

    Isihara, T. | Tema Archiv | 1983



    Roca electrification out to tender

    British Library Online Contents | 2008