Boosted Aquila Arithmetic Optimization Algorithm for multi-level thresholding image segmentation

The traditional threshold methods used for image segmentation are effective for bi-level thresholds. In the case of complex images that contain many objects or color images, the computational complexity is significantly elevated. Multi-level threshold methods for the segmentation of color images can be seen as a complicated optimization problem. In this paper, an improved version of the Arithmetic Optimization Algorithm, called AOAa, is proposed based on the efficient search operators of Aquila Optimizer to obtain optimal threshold values in various levels of color and gray images. Otsu and Kapur's entropy methods are used in this study as objective functions. Experiments were conducted on 16 benchmark images; COVID-19, color, and gray. The results are analyzed regarding the fitness function, peak signal-to-noise ratio (PSNR), and structural index similarity (SSIM). The obtained results showed that the proposed method got better results than several other well-established methods.