Multilevel Thresholding Using Chaos Optimization and Differential Evolution Algorithm
Yuhuan Chen and Chenfu Yi
Pages 206-217 (12)
Multilevel thresholding usually is much computationally exhaustive in the process of searching the optimal
thresholds. In order to improve computational efficiency, this paper presents an image segmentation method by using the
chaos optimization algorithm (COA), which is incorporated into differential evolution (DE). The stochastic property and
space ergodicity of chaos mapping are utilized to enlarge the search range and to explore a huge search space. Additionally,
to find the optimal thresholds, the differential evolution with chaos optimization algorithm (DECOA) is considered by
using the objective model based on the maximum entropy criterion. The presented segmentation method has been
simulated on six standard test images and compared with the canonical DE and some other classic optimization
algorithms. Experimental results show that the presented DECOA algorithm has much faster convergence speed than
those of some existing methods. Furthermore, this algorithm can get superior segmentation performance of the feasibility
Chaos optimization, differential evolution, image segmentation, maximum entropy, multilevel thresholding.
Research center for Biomedical and Information Technology, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China.