View in EDS

GAAOA-Lévy: a hybrid metaheuristic for optimized multilevel thresholding in image segmentation

Saved in:
Bibliographic Details
Title: GAAOA-Lévy: a hybrid metaheuristic for optimized multilevel thresholding in image segmentation
Authors: Eman Mahmoud, Salem Alkhalaf, Tomonobu Senjyu, Masahiro Furukakoi, Ashraf Hemeida, Ghada Abozaid
Source: Scientific Reports, Vol 15, Iss 1, Pp 1-24 (2025)
Publisher Information: Nature Portfolio, 2025.
Publication Year: 2025
Collection: LCC:Medicine
LCC:Science
Subject Terms: Image segmentation, Genetic algorithm (GA), Archimedes optimization algorithm (AOA), Multilevel thresholding, Medicine, Science
Description: Abstract Image segmentation is a critical task in image processing with applications in various domains, including industry and medicine. However, multilevel thresholding, a widely used segmentation technique, suffers from high computational complexity due to the exhaustive search for optimal thresholds. This paper addresses this challenge by proposing a hybrid Genetic Algorithm-Archimedes Optimization Algorithm (GAAOA), further enhanced with a Lévy flight function (GAAOA-Lévy), to improve efficiency and accuracy in multilevel thresholding. The integration of GA’s crossover mechanism strengthens local search capabilities, leading to optimal segmentation with fewer iterations. The proposed algorithm is evaluated using standard benchmark images and compared against well-known optimization techniques. Experimental results demonstrate that GAAOA-Lévy outperforms existing methods in terms of Peak Signal-to-Noise Ratio (PSNR), computational efficiency, and convergence speed, particularly excelling in three-level thresholding while reducing computational costs for higher thresholds.
Document Type: article
File Description: electronic resource
Language: English
ISSN: 2045-2322
Relation: https://doaj.org/toc/2045-2322
DOI: 10.1038/s41598-025-12142-z
Access URL: https://doaj.org/article/3a0393559ced4fa4888a44bfa3bac12b
Accession Number: edsdoj.3a0393559ced4fa4888a44bfa3bac12b
Database: Directory of Open Access Journals
Description
Abstract:Abstract Image segmentation is a critical task in image processing with applications in various domains, including industry and medicine. However, multilevel thresholding, a widely used segmentation technique, suffers from high computational complexity due to the exhaustive search for optimal thresholds. This paper addresses this challenge by proposing a hybrid Genetic Algorithm-Archimedes Optimization Algorithm (GAAOA), further enhanced with a Lévy flight function (GAAOA-Lévy), to improve efficiency and accuracy in multilevel thresholding. The integration of GA’s crossover mechanism strengthens local search capabilities, leading to optimal segmentation with fewer iterations. The proposed algorithm is evaluated using standard benchmark images and compared against well-known optimization techniques. Experimental results demonstrate that GAAOA-Lévy outperforms existing methods in terms of Peak Signal-to-Noise Ratio (PSNR), computational efficiency, and convergence speed, particularly excelling in three-level thresholding while reducing computational costs for higher thresholds.
ISSN:20452322
DOI:10.1038/s41598-025-12142-z