Multilevel thresholding is a simple and powerful image segmentation method that has received widespread attention in recent years. However, the accuracy and stability of thresholding techniques are still not ideal and cannot meet the needs of engineering problems. Therefore, this paper presents a memetic algorithm of dragonfly algorithm (DA) and differential evolution (DE) for color image segmentation, which is known as improved DA (IDA). On the one hand, DA algorithm has a satisfied capability of avoiding convergence to the local optimum, thus it is served as a global search technique. On the other hand, the DE algorithm is adopted as a local search technique, which can increase the precision of solutions. In this paper, two thresholding techniques, namely, Otsu and minimum cross entropy (MCE) methods are used to determine the optimal threshold values. In order to evaluate the performance of the proposed method, we conduct a series of experiments on color images from the Berkeley database and the results are compared with five stateof-the-art meta-heuristic algorithms. Besides, a non-parametric Wilcoxon's rank sum test is also included for statistical analysis. From the experimental results, it is found that IDA-based method outperforms other compared methods in terms of average fitness values, standard deviation, peak signal to noise ratio, structural similarity index, and feature similarity index. The promising results indicate that the application of IDA-based thresholding technique is potential and meaningful.
Chimp optimization algorithm (ChOA) is a recently proposed metaheuristic. Interestingly, it simulates the social status relationship and hunting behavior of chimps. Due to the more flexible and complex application fields, researchers have higher requirements for native algorithms. In this paper, an enhanced chimp optimization algorithm (EChOA) is proposed to improve the accuracy of solutions. First, the highly disruptive polynomial mutation is used to initialize the population, which provides the foundation for global search. Next, Spearman’s rank correlation coefficient of the chimps with the lowest social status is calculated with respect to the leader chimp. To reduce the probability of falling into the local optimum, the beetle antennae operator is used to improve the less fit chimps while gaining visual capability. Three strategies enhance the exploration and exploitation of the native algorithm. To verify the function optimization performance, EChOA is comprehensively analyzed on 12 classical benchmark functions and 15 CEC2017 benchmark functions. Besides, the practicability of EChOA is also highlighted by three engineering design problems and training multilayer perceptron. Compared with ChOA and five state-of-the-art algorithms, the statistical results show that EChOA has strong competitive capabilities and promising prospects.
Multithreshold segmentation is an indispensable part of modern image processing. Color images contain more information than gray images, therefore RGB multi-thresholding segmentation techniques have been drawn much attention during recent years. Multiverse optimization (MVO) algorithm has a strong advantage in finding the optimal solution of three channels for RGB. In this paper, an MVO algorithm based on Lévy flight (LMVO) is proposed. Lévy flight is an efficient strategy which can not only increase the population diversity to prevent premature convergence but also improve the ability to jump out of the local optimum. Therefore, LMVO conduces to achieve a better balance between exploration and exploitation of MVO, so that it is faster and more robust than MVO and avoids premature convergence. Further LMVO algorithm is compared with the other eight famous meta-heuristics algorithms, by maximizing the objective function of Kapur's entropy method or of Otsu method to determine the optimal threshold. The maximum objective function, peak signal-to-noise ratio (PSNR), feature similarity index (FSIM), structural similarity index (SSIM), CPU calculation time, optimal threshold value, and Wilcoxon's rank-sum test are used to evaluate the quality of the segmented image. The experimental results show that this method has obvious advantages in terms of objective function value, image quality measurement, convergence performance, and robustness.
In order to realize the multilevel thresholding segmentation of color satellite images, a multistrategy emperor penguin optimizer (called MSEPO) is proposed to find the optimal threshold values for three channels of RGB images. Masi entropy is utilized as the objective function. Meanwhile, three strategies are introduced, namely highly disruptive polynomial mutation, Levy flight, and thermal exchange operator. Through these, the MSEPO is able to properly balance the exploration and exploitation mechanisms. Moreover, the convergence, accuracy and stability performance have been significantly enhanced. Tests are carried out on color Berkeley images and color satellite images at various threshold levels. The experimental results show that the proposed method achieves higher Peak Signal to Noise Ratio (PSNR), higher Structural Similarity Index (SSIM), higher Feature Similarity Index (FSIM), and shorter CPU time than seven state-ofthe-art optimization techniques. To present in a comprehensive manner, the computational complexity has also been analyzed in terms of time and space complexity. Wilcoxon rank sum test and Friedman test are also applied to statistical analysis. To sum up, MSEPO algorithm has achieved significant improvement and superior performance. What's more, the proposed technique is more suitable for high-dimensional segmentation of complex satellite images.INDEX TERMS Multilevel thresholding, satellite image segmentation, Masi entropy, emperor penguin optimizer, thermal exchange operator, multi-strategy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.