Abstract
One propounds two methods of image contour segmentation based on the tonotopy property of Kohonen’s self-organizing maps (SOM). The first method consists in quantizing the set of gray levels of the image by making use of a one dimensional SOM, then detecting the spatial discontinuities of the involved mapping of quantized gray levels onto map’s cells. The same method brings color image segmentation when quantizing the color values of the image by making use of a three dimensional SOM. This first method is multiresolution acording to gray level or color accuracy. Examples of gray level and color image segmentations illustrate its outcome.
The second method presented consists in quantizing the set of spatial and gray level pixel coordinates by a three dimensional SOM. This quantization brings a “region-like” pixel clustering out of which one computes contour segmentation of the image. This method is monoresolution and yelds contour segmentations less noisy than the one obtained by the first method. This second method is illultrated by an example of gray level image segmentation.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
T. Kohonen, “Self-organization and associative memory”, Springer-Verlag Berlin, 1984.
T. Kohonen, “The self-organizing feature map”; proceedings of the I.E.E.E., vol. 78, n. 9, September 1990.
R. Natowicz, R. Sokol, “Self-organizing feature maps for image segmentation”, I.W.A.N.N.’93, lecture notes in computer science, vol. 686, Springer-Verlag, 1993.
R. Natowicz, “Segmentation d’images par cartes de Kohonen”, colloque Gretsi, Juan les Pins, 1993.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1995 Springer-Verlag/Wien
About this paper
Cite this paper
Natowicz, R., Bergen, L., Gas, B. (1995). Kohonen’s Maps for Contour and “Region-Like” Segmentation of Gray Level and Color Images. In: Artificial Neural Nets and Genetic Algorithms. Springer, Vienna. https://doi.org/10.1007/978-3-7091-7535-4_94
Download citation
DOI: https://doi.org/10.1007/978-3-7091-7535-4_94
Publisher Name: Springer, Vienna
Print ISBN: 978-3-211-82692-8
Online ISBN: 978-3-7091-7535-4
eBook Packages: Springer Book Archive