Electronic Letters on Computer Vision and Image Analysis 7(3):93-100, 2008
Complex networks :
application for texture characterization and classification
T. Chalumeau∗ and L. da F. Costa+ and O. Laligant∗ and F. Meriaudeau∗
∗ Universite de Bourgogne, Le2i Laboratory, 12 rue de la Fonderie 71200, Le Creusot, France
+ Universidade de Sao Paulo, IFSC- Av. Trabalhador Sao-Carlense, 400 Centro, Sao Carlos (SP), Brasil
Received 16th May 2008; accepted 12th March 2009
This article describes a new method and approach of texture characterization. Using complex network
representation of an image, classical and derived (hierarchical) measurements, we present how to have good
performance in texture classification. Image is represented by a complex networks : one pixel as a node.
Node degree and clustering coefficient, using with traditional and extended hierarchical measurements, are
used to characterize ”organization” of textures.
Key Words: Image processing, texture analysis, complex networks.
Texture analysis have important role in numerous application of image processing. Many different approaches
to texture analysis have been proposed. Among the most widely used texture measures are those derived from
gray level co-occurence matrices or difference histograms, ”texture energy” measures obtained by local linear
transforms, and features based on multi-channel Gabor filtering or Markov random field model [1, 2].
Introduced recently [3, 4, 5], complex networks can be adapted to represent the relation and characteriza-
tion between elements and become appropriate to characterize picture pattern. It is possible to represent an
image as a complex network and used tools from texture networks theory to characterize the created image:
segmentation , texture analysis .
This paper overviews our approach, presents in the first part complex networks and image representation, in
the second part methods that were used for comparison. The third part exposes complex networks method’s
results, with the efficie