An image segmentation algorithm based on fuzzy clustering and genetic algorithms with a new distance

  • Authors:

  • F. Hachouf;A. Zeggari;Z. Ahmed Seghir

  • Affiliations:

  • Department of electronics, Automatic and Robotic laboratory, Mentouri Constantine University, Constantine, Algeria;Department of Exact Science and Technology, Academic Center of Tebessa, Algeria;Department of Exact Science and Technology, Academic Center of LARBI BEN M'HIDI, Oum el bouaghi, Algeria

  • Venue:
  • EC'05 Proceedings of the 6th WSEAS international conference on Evolutionary computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes a new GA-clustering algorithm for image segmentation. We combine the classical fuzzy c-means algorithm (FCM) with a genetic algorithm, and we modify the distance function in FCM for taking into account the spatial information and the color of a pixel. Image segmentation is treated as an unsupervised classification which is optimised by a genetic algorithm. The idea is to choose several configurations of initial centres and ton code chromosomes by the membership degrees of pixels to the clusters. The new proposed distance yield uniform regions while respecting the quality of segmentation.