mk:annes95

Summary

Adaptation of Morphological Masks by Genetic Algorithms. Mario Köppen and Martin Teunis. In Proceedings ANNES'95, Dunedin, New Zealand, pages 398-403, 1995.

Abstract

Sequences of basic operations of mathematical morphology can perform high-level image processing tasks. Morphological operations make use of a so-called structuring element, which should be adapted to improve the performance of certain tasks. A serious problem arises from the fact that greyscale morphology influences the grey-value distribution in the image in an unpredictable manner. This problem was not addressed in literature so far. This paper introduces a new fitness function that overcomes this disadvantage and offers new possibilities for the adaptation of structuring elements in mathematical morphology. The new approach is based on the comparison of performing the same operations sequence on the starting image and a scattered variant of this image. The more this two images differ the better the fitness of the structuring element. Also a method is introduced to estimate the histogram of the operated scattered image. A result is given for the application of this method in texture analysis.

Bibtex entry

@INPROCEEDINGS { mk:annes95,
    ABSTRACT = { Sequences of basic operations of mathematical morphology can perform high-level image processing tasks. Morphological operations make use of a so-called structuring element, which should be adapted to improve the performance of certain tasks. A serious problem arises from the fact that greyscale morphology influences the grey-value distribution in the image in an unpredictable manner. This problem was not addressed in literature so far. This paper introduces a new fitness function that overcomes this disadvantage and offers new possibilities for the adaptation of structuring elements in mathematical morphology. The new approach is based on the comparison of performing the same operations sequence on the starting image and a scattered variant of this image. The more this two images differ the better the fitness of the structuring element. Also a method is introduced to estimate the histogram of the operated scattered image. A result is given for the application of this method in texture analysis. },
    AUTHOR = { Mario Köppen and Martin Teunis },
    BOOKTITLE = { Proceedings ANNES'95, Dunedin, New Zealand },
    MODIFIED = { 2008-02-28 17:03:30 +0900 },
    HASABSTRACT = { Yes },
    PAGES = { 398--403 },
    PDF = { annes95.pdf },
    TITLE = { Adaptation of Morphological Masks by Genetic Algorithms },
    YEAR = { 1995 },
}