Improved grasshopper optimization and modified moth–flame optimization (IGH-MMFO) algorithm for identifying optimal threshold in image segmentation

• Published in: Neural computing & applications Vol. 37; no. 22; pp. 18609 - 18631
• Main Authors: Manikandan, K., Sudhakar, B.
• Format: Journal Article
• Language: English
• Published: London Springer London 01.08.2025; Springer Nature B.V
• Subjects: Algorithms; Ant colony optimization; Artificial Intelligence; Bees; Butterflies & moths; Comparative analysis; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Entropy; Exploitation; Genetic algorithms; Grasshoppers; Image Processing and Computer Vision; Image segmentation; Mold; Optimization algorithms; Optimization models; Optimization techniques; Original Article; Probability and Statistics in Computer Science; Signal to noise ratio; Thresholds; Grasshopper optimization; Moth–flame optimization; Optimization algorithms; Image segmentation; Multilevel thresholding
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-025-11391-3

Multilevel thresholding is a common procedure followed in image segmentation. Several regions in an image are divided into multiple parts based on the threshold values. However, manually determining optimal thresholds for image segmentation is time-consuming and requires exhaustive analysis. Thus, optimization algorithms have been incorporated in recent times to select the optimal thresholds based on optimal solutions. However, the conventional optimization models have limitations in their search space and dimensions. Thus, in this research work, a hybrid optimization algorithm is presented for finding the optimal threshold in image segmentation. Improved grasshopper optimization and Modified Moth–flame optimization (IGH-MMFO) algorithms are comprised in the hybrid approach, which identifies the optimal threshold. The improved grasshopper optimization (IGH) model includes Levy flight to enhance the search diversity and strength of the conventional model. Additionally, for enhanced searchability of IGH, modified moth–flame optimization is incorporated in the proposed work. Better performance of proposed IGH-MMFO is validated through metrics like peak signal-to-noise ratio, standard deviation, mean, and structural similarity index metrics over conventional and hybrid optimization algorithms like grey wolf optimization, bee foraging algorithm, Ant colony optimization (ACO), salp swarm optimization, hybrid Salp Swarm Ant Colony Optimization (SS-ACO), hybrid Improved flower pollination and bee foraging algorithm.