Pair correlation integral for fractal characterization of three-dimensional histograms from color images
Journal Title: Lecture Notes in Computer Science
Article Title: Pair correlation integral for fractal characterization of three-dimensional histograms from color images
Volume: LNCS 5099
First Page: 200
Last Page: 208
Issue Cover Date: 2008
Authors: Julien Chauveau1, David Rousseau1 and François Chapeau-Blondeau1
- Laboratoire d’Ingénierie des Systèmes Automatisés (LISA),
Université d’Angers, 62 avenue Notre Dame du Lac, 49000 Angers, France
Abstract: The pair correlation integral is used to assess the intrinsic dimensionality of the three-dimensional histogram of RGB color images. For application in the bounded colorimetric cube, this correlation measure is first calibrated on color histograms of reference constructed with integer dimensionality. The measure is then applied to natural color images. The results show that their color histogram tends to display a self-similar structure with noninteger fractal dimension. Such a fractal organization in the colorimetric space can have relevance for image segmentation or classification, or other areas of color image processing.
