Febri Liantoni, Kartika Candra Kirana
Ant Colony Optimization (ACO) is widely used for optimization tasks, inspired by the foraging behaviour of ants. However, the original ACO algorithm suffers from an unbalanced random distribution of ants, which can impede the discovery of optimal paths. To address this issue, this research proposes an adaptive version of the ACO algorithm that improves edge detection by distributing ants based on gradient analysis. Specifically, ants are adaptively allocated according to the gradient ratio of each image pixel, replacing arbitrary placement with a pixel-specific approach. This technique leverages image gradients to ensure a more precise distribution of ants along the edges. Experimental results demonstrate that the adaptive ACO algorithm outperforms the standard ACO algorithm by delivering better image features for edge detection and achieving more precise, well-defined edges. Furthermore, the strategic placement of ants based on gradient values significantly enhances edge detection accuracy compared to random distribution. The Adaptive ACO method also demonstrated superior optimization in edge detection processes, as evidenced by its highest average PSNR value of 11.714, compared to 9.701 for Sobel edges and 9.883 for conventional ACO. These results highlight the Adaptive ACO method's ability to not only improve edge detection accuracy but also streamline the process for faster and more reliable path discovery. © 2025 IEEE.
Sebelas Maret University, Department of Computer and Information Education, Surakarta, Indonesia; Universitas Negeri Malang, Department of Electrical Engineering, Malang, Indonesia
