A modified genetic algorithm for feature clustering

  title={A modified genetic algorithm for feature clustering},
  author={Miki Aoyagi and Kumiko Tsuji},
A new genetic algorithm (GA) based on feature clustering with an energy function is proposed for obtaining optimal segmentation. In the proposed algorithm, which we call the modified genetic algorithm (MGA), the length of each genome is the number of features and each individual (genome) represents one assignment of the input-features to labels. The energy function, which is used as a fitness function for our GA, consists of local competitive and cooperative interactions among features and a… CONTINUE READING
1 Citations
4 References
Similar Papers


Publications citing this paper.


Publications referenced by this paper.
Showing 1-4 of 4 references

A modified genetic algorithm for image segmentation based on feature clustering

  • M. Aoyagi, K. Tsuji
  • to appear in International Symposium on…
  • 2004
3 Excerpts

Similar Papers

Loading similar papers…