@InProceedings{MoraesHaer:2007:MéHiRe,
               author = "Moraes, Denis Altieri de Oliveira and Haertel, Vitor Francisco de 
                         Ara{\'u}jo",
          affiliation = "{Universidade Federal do Rio Grande do Sul (UFRGS). Centro 
                         Estadual de Pesquisas em Sensoriamento Remoto e Metereologia 
                         (CEPSRM).} and {Universidade Federal do Rio Grande do Sul (UFRGS). 
                         Centro Estadual de Pesquisas em Sensoriamento Remoto e 
                         Metereologia (CEPSRM).}",
                title = "M{\'e}todos hier{\'a}rquicos para redu{\c{c}}{\~a}o de 
                         dimens{\~o}es e classifica{\c{c}}{\~a}o de imagens AVIRIS",
            booktitle = "Anais...",
                 year = "2007",
               editor = "Epiphanio, Jos{\'e} Carlos Neves and Galv{\~a}o, L{\^e}nio 
                         Soares and Fonseca, Leila Maria Garcia",
                pages = "6481--6488",
         organization = "Simp{\'o}sio Brasileiro de Sensoriamento Remoto, 13. (SBSR).",
            publisher = "Instituto Nacional de Pesquisas Espaciais (INPE)",
              address = "S{\~a}o Jos{\'e} dos Campos",
             keywords = "Reconhecimento de padr{\~o}es, imagem hiperespectral, 
                         {\'a}rvores de decis{\~a}o, dist{\^a}ncia de Bhattacharyya, 
                         sele{\c{c}}{\~a}o de vari{\'a}veis, extra{\c{c}}{\~a}o de 
                         vari{\'a}veis.",
             abstract = "In this paper we investigate the performance of a tree structured 
                         classifier in the labeling of high dimensional image data. The aim 
                         is to mitigate the effects of the Hughes phenomenon in 
                         hyperspectral image data classification. The use of a multi-stage 
                         classifier, analyzing a sub-set of classes at each stage rather 
                         than the full set at once, allows for a more efficient way to 
                         reduce the data dimensionality. The extracted features can then be 
                         selected in order to maximize the discrimination between the 
                         sub-set of classes under consideration. In a binary tree approach, 
                         only two classes are considered at each node, allowing for the 
                         implementation of the Bhattacharyya distance as a criterion for 
                         feature extraction at each tree node. Experiments are performed 
                         using AVIRIS image data. The performance of the proposed 
                         methodology is compared against the more traditional methods for 
                         feature selection and extraction.",
  conference-location = "Florian{\'o}polis",
      conference-year = "21-26 abr. 2007",
                 isbn = "978-85-17-00031-7",
             language = "pt",
         organisation = "Instituto Nacional de Pesquisas Espaciais (INPE)",
                  ibi = "dpi.inpe.br/sbsr@80/2006/11.14.17.30",
                  url = "http://urlib.net/ibi/dpi.inpe.br/sbsr@80/2006/11.14.17.30",
           targetfile = "6481-6488.pdf",
                 type = "Sensoriamento Remoto Hiperespectral",
        urlaccessdate = "15 jun. 2024"
}