Close

%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2016/09.06.18.06
%2 sid.inpe.br/sibgrapi/2016/09.06.18.06.25
%T Ranking Eigenfaces Through Adaboost and Perceptron Ensembles
%D 2016
%A Filisbino, Tiene Andre,
%A Giraldi, Gilson Antonio,
%A Thomaz, Carlos Eduardo,
%@affiliation Laboratorio Nacional de Comnputação Científica
%@affiliation Laboratorio Nacional de Comnputação Científica
%@affiliation Centro Universitário da FEI
%E Aliaga, Daniel G.,
%E Davis, Larry S.,
%E Farias, Ricardo C.,
%E Fernandes, Leandro A. F.,
%E Gibson, Stuart J.,
%E Giraldi, Gilson A.,
%E Gois, João Paulo,
%E Maciel, Anderson,
%E Menotti, David,
%E Miranda, Paulo A. V.,
%E Musse, Soraia,
%E Namikawa, Laercio,
%E Pamplona, Mauricio,
%E Papa, João Paulo,
%E Santos, Jefersson dos,
%E Schwartz, William Robson,
%E Thomaz, Carlos E.,
%B Conference on Graphics, Patterns and Images, 29 (SIBGRAPI)
%C São José dos Campos, SP, Brazil
%8 4-7 Oct. 2016
%I Sociedade Brasileira de Computação
%J Porto Alegre
%S Proceedings
%K Ranking PCA Components, Separating Hyperplanes, Perceptron, AdaBoost, Face Image Analysis.
%X The fact that principal component analysis (PCA) does not necessarily represent important discriminant directions to separate sample groups motivates the development of the multi-class discriminant principal component analysis (MDPCA). This technique addresses the problem of ranking face features in N-class problems computed by PCA components (eigenfaces). Given a database, the MDPCA builds a linear support vector machine (SVM) ensemble to get the separating hyperplanes that are combined through an AdaBoost technique to determine the discriminant contribution of each PCA feature. In this paper, we follow the MDPCA methodology but we replace the SVM by the linear perceptron as the basic learner in the AdaBoost approach. In the computational experiments we compare the obtained technique, called MDPCA-Perceptron, with the PCA and the original MDPCA through facial expression experiments. Our computational results have shown that the principal components selected by the MDPCA-Perceptron allow competitive recognition rates in lower dimensional spaces with promising results for reconstruction tasks as well.
%@language en
%3 REAPE2.pdf


Close