%0 Thesis
%4 ime.usp.br/hirata/1997/09.02.16.38
%2 dpi.inpe.br/banon/2001/02.23.19.30.12
%T Segmentação de Imagens por Morfologia Matemática
%J Image segmentation by Mathematical Morphology
%D 1997
%8 1997-03
%9 Masters Thesis
%A Hirata Junior, Roberto,
%I Instituto de Matemática e Estatística - USP
%C São Paulo
%K mathematical morphology, image segmentation, image analysis, fast algorithms.
%X This work shows what segmentation is and how to segment an image using Mathematical Morphology (\MM). It begins by introducing the subject segmentation by the classical approach and then by the morphological one. A short introduction to \MM is given and then two morphological approaches are given for segmentation: an heuristic and a non-heuristic one. The equivalence between the classical and the morphological approaches is shown and several illustrative examples of the morphological approach are given. The work also shows some important fast algorithms which simulate some of the operators used in the segmentations and how to use them to segment images. Finally it shows how to use the ideas behind the fast algorithms to implement efficiently the elementary operations and operators of \MM so other complex operators can be implemented more efficiently than before. Some experimental results are shown to verify this assumption.
%@language pt
%3 vfinal.pdf
%O This work has been supported by ProTeM-CC/CNPq through the AnIMoMat project, contract 680067/94-9.