Hybrid Multiverse Optimization Algorithm With Gravitational Search Algorithm for Multithreshold Color Image Segmentation

Abstract

Multithreshold segmentation of color images is an important part of modern image processing. Multithreshold segmentation techniques that use the ideas of optimization in mathematics to process the information contained in the three channels of RGB images have received much attention in recent years. In this paper, a hybrid algorithm of multiverse optimization algorithm with a gravitational search algorithm (GSMVO) is proposed. On the one hand, the gravitational mechanism among individuals in the universe enables particles to share information and move toward particles with large mass (the optimal solution), thus optimizing local optimization. On the other hand, the problem of a slow interpopulation update is improved. On the premise of improving the optimization precision, the convergence speed of the whole system is accelerated and the robustness can be improved. The time complexity of a proposed hybrid algorithm is analyzed. The quality of a segmented image is evaluated by using peak signal-to-noise ratio (PSNR), feature similarity index (FSIM), structural similarity index (SSIM), and Wilcoxon test. The experimental results illustrate that the proposed method has obvious advantages in objective function value, image quality measurement, convergence performance, and robustness.

Keywords

• Multiverse optimization algorithm
• gravitational search algorithm
• multithreshold segmentation
• Kapur’s entropy
• Otsu method