An efficient multilevel color image thresholding based on modified whale optimization algorithm

Color image segmentation is a vital preprocessing stage in various image processing applications. In thresholdbased segmentation, the success of the image segmentation depends mainly on the optimal selection of thresholds. The selection of threshold values for multilevel thresholding is indeed a time-consuming process compared to bi-level thresholding. The issue of optimal threshold selection is formulated as an optimization problem in the case of color image segmentation using multilevel thresholding. To optimize the threshold selection for a multilevel color image thresholding, a modified whale optimization algorithm (MWOA) is proposed in this paper. The Otsu's and Kapur's functions have been used in the proposed strategy as a fitness function that can be maximized by MWOA. In the MWOA, the position of the whales is controlled by adapting the cosine function during optimization process. Further, the movements of search agents are regulated during the search process by introducing the correction factors in position updation. These changes incorporated in the MWOA provide a proper balance between the phases of exploration and exploitation and also avoid local optima problem. The performance of the MWOA is evaluated quantitatively and qualitatively based on the best fitness values in terms of PSNR, SSIM, and FSIM, further CPU computing time and Wilcoxon test. The experimental outcomes show that the proposed multilevel optimal color image thresholding using MWOA algorithm yields better performance results in terms of image quality, feature conservation, and convergence rate with less CPU computing time than other state-of-the-art algorithms.

Automated summary

This paper proposes a modified whale optimization algorithm for optimizing the selection of multilevel color image thresholds. The algorithm achieves a proper balance between exploration and exploitation phases, thus avoiding the issue of local optima.