An improved ant colony algorithm for fuzzy clustering in image segmentation
- 31 January 2007
- journal article
- Published by Elsevier BV in Neurocomputing
- Vol. 70 (4-6), 665-671
- https://doi.org/10.1016/j.neucom.2006.10.022
Abstract
Ant colony algorithm (ACA), inspired by the food-searching behavior of ants, is an evolutionary algorithm and performs well in discrete optimization. In this paper, it is used for fuzzy clustering in image segmentation. Three features such as gray value, gradient and neighborhood of the pixels, are extracted for the searching and clustering process. Unexpectedly, tests show that it is time consuming when dealing with the vast image data. In view of this drawback, improvements have been made by initializing the clustering centers and enhancing the heuristic function to accelerate the searching process. Experiments and comparisons are done to show that the improved ACA-based image segmentation is an efficient and effective approach.Keywords
This publication has 17 references indexed in Scilit:
- Efficient region segmentation on compressed gray images using quadtree and shading representationPattern Recognition, 2004
- Quadtree-based genetic algorithm and its applications to computer visionPattern Recognition, 2004
- FACE: fast active-contour curvature-based evolutionSignal Processing: Image Communication, 2004
- Robust and automated unimodal histogram thresholding and potential applicationsPattern Recognition, 2004
- Fast accurate fuzzy clustering through data reductionIEEE Transactions on Fuzzy Systems, 2003
- Image processing with neural networks—a reviewPattern Recognition, 2002
- Watershed-Based Segmentation and Region MergingComputer Vision and Image Understanding, 2000
- Comparison of Edge DetectorsComputer Vision and Image Understanding, 1998
- A new minimum variance region growing algorithm for image segmentationPattern Recognition Letters, 1997
- Real-time corner detection algorithm for motion estimationImage and Vision Computing, 1995