An APO Algorithm Based on Taguchi Methods and Its Application in Multi-Level Image Segmentation

Multilevel image segmentation is a critical task in image analysis, which imposes high requirements on the global search capability and convergence efficiency of segmentation algorithms. In this paper, an improved Artificial Protozoa Optimization algorithm, termed the two-stage Taguchi-assisted Gaussian–Lévy Artificial Protozoa Optimization (TGAPO) algorithm, is proposed and applied to multilevel image segmentation. The proposed algorithm adopts a two-stage evolutionary mechanism. In the first stage, Gaussian perturbation is introduced to enhance local search capability; in the second stage, Lévy flight is incorporated to expand the global search range; and finally, the Taguchi strategy is employed to further refine the optimal solution. Consequently, the global optimization performance and robustness of the algorithm are significantly improved. To evaluate the effectiveness of the proposed TGAPO algorithm, comparative experiments are conducted with representative optimization algorithms, including the Grey Wolf Optimizer (GWO) and Particle Swarm Optimization (PSO), in the context of multilevel image segmentation. The segmentation quality is assessed using the minimum cross-entropy function as the performance metric. Experimental results demonstrate that the TGAPO algorithm outperforms the comparison algorithms in terms of segmentation accuracy and convergence speed, and exhibits superior stability in high-threshold segmentation tasks. Furthermore, the proposed method achieves excellent multi-threshold segmentation performance for color images and shows strong potential for practical applications.