Title :
Differential evolution and particle swarm optimization in fuzzy c-partition
Author_Institution :
Dept. of Comput. Sci., Univ. of M´sila, M´sila, Algeria
Abstract :
The fuzzy c-partition entropy approach for threshold selection is one of the best image thresholding techniques, but its complexity increases with the number of thresholds. In this paper, the selection of thresholds (fuzzy parameters) was seen as an optimization problem and solved using particle swarm optimization (PSO) and differential evolution (DE) algorithms. The proposed fast approaches have been tested on many images. For example, the processing time of four-level thresholding using both PSO and DE is reduced to less than 0.4s. PSO and DE show equal performance when the number of thresholds is small. When the number of thresholds is greater, PSO algorithm performs better than DE in terms of precision, robustness and execution time.
Keywords :
evolutionary computation; feature selection; fuzzy set theory; image segmentation; particle swarm optimisation; DE; PSO; differential evolution; fuzzy c-partition entropy; image thresholding; particle swarm optimization; threshold selection; Artificial intelligence; Bismuth; Lakes; Differential Evolution Algorithm; Entropy; Fuzzy c-partition; Histograms; Image segmentation; Optimization; Particle swarm optimization; Partitioning algorithms; Pattern recognition; Thresholding;
Conference_Titel :
Multimedia Computing and Systems (ICMCS), 2014 International Conference on
Conference_Location :
Marrakech
Print_ISBN :
978-1-4799-3823-0
DOI :
10.1109/ICMCS.2014.6911136