ISTOA: An Improved Sooty Tern Optimization Algorithm for Multilevel Threshold Image Segmentation

This paper introduces an improved version of well-known Sooty Tern Optimization Algorithm (STOA). The improved version combines Opposition based learning (OBL) to introduce the Improved Sooty Tern Optimization Algorithm (ISTOA). The OBL strategy increases population diversity and avoids falling into local solutions. The efficiency of the proposed ISTOA is verified on multilevel threshold segmentation based on the objective functions of Kapur, and its performance is compared with the original algorithm and another metaheuristic algorithm. Experimental results reveal that the proposed ISTOA outperforms other algorithms in terms of fitness, peak signal-to-noise ratio, structural similarity, and segmentation findings.