An Improved Beluga Whale Optimization Algorithm by Collaborative Strategies for Multi-Threshold Image Segmentation

Multi-threshold segmentation techniques usually rely on exhaustive search, leading to an exponential growth in computational complexity with the increase in the number of thresholds. The combination of swarm intelligence and image segmentation is a promising approach to improve segmentation efficiency and reduce computational cost. Thus, this paper then proposes an Improved Beluga Whale Optimization (IBWO) algorithm incorporating collaborative strategies for multi-threshold image segmentation. As for the proposed IBWO algorithm, the following collaborative strategies are utilized. The logistic chaos mapping strategy can increase the randomness of the initial beluga population, the dynamic nonlinear equilibrium factor strategy can better balance the exploratory and exploitation phases, and the dynamic pinhole imaging strategy can increase the diversity of the population and prevent the population from falling into a local optimum. The objective function is minimum cross entropy, and experiments use benchmark images at different threshold levels to evaluate the method's performance and compare the IBWO algorithm with several other classical algorithms. The experimental results show that the proposed IBWO algorithm outperforms the comparative methods in all three image quality metrics, which are PSNR, SSIM, and FSIM.