Refined Coati Optimizer with Chaotic Systems for Optimization in Image Processing Applications

The Coati Optimization Algorithm (COA) is a promising metaheuristic inspired by coati hunting behaviors, demonstrating strong performance in complex optimization tasks. However, COA encounters challenges with convergence speed and is susceptible to getting trapped in local optima. This paper introduces the Refined Coati Optimizer (R-CO), an enhanced COA variant incorporating chaotic systems to improve exploration and exploitation phases. By embedding chaotic maps, R-CO increases search diversity, accelerates convergence, and reduces the likelihood of premature convergence. Extensive testing on 23 benchmark functions—covering unimodal, multimodal, and fixed-dimension multimodal problems—confirms the algorithm’s robust performance. $\mathrm{R}-\mathrm{CO}$ is also applied to an image segmentation task, where it outperforms the original COA and other leading metaheuristic algorithms. Results highlight R-CO’s enhanced optimization capabilities, positioning it as a valuable tool for theoretical and applied optimization challenges.