Imagen de Portada

Improved Weighted Chimp Optimization Algorithm Based on Fitness–Distance Balance for Multilevel Thresholding Image Segmentation

Guardado en:
Publicado en:Symmetry vol. 17, no. 7 (2025), p. 1066-1103
Autor principal: Günay, Yılmaz Asuman
Otros Autores: Samoua, Alsamoua
Publicado:
MDPI AG
Materias:
Similarity
Performance evaluation
Image segmentation
Computer vision
Color imagery
Fitness
Statistical tests
Multilevel
Task complexity
Optimization
Statistical methods
Solution space
Search algorithms
Image processing
Optimization algorithms
Entropy
Heuristic methods
Signal to noise ratio
Acceso en línea:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Resumen:Multilevel thresholding image segmentation plays a crucial role in various image processing applications. However, achieving optimal segmentation results often poses challenges due to the intricate nature of images. In this study, a novel metaheuristic search algorithm named Weighted Chimp Optimization Algorithm with Fitness–Distance Balance (WChOA-FDB) is developed. The algorithm integrates the concept of Fitness–Distance Balance (FDB) to ensure balanced exploration and exploitation of the solution space, thus enhancing convergence speed and solution quality. Moreover, WChOA-FDB incorporates weighted Chimp Optimization Algorithm techniques to further improve its performance in handling multilevel thresholding challenges. Experimental studies were conducted to test and verify the developed method. The algorithm’s performance was evaluated using 10 benchmark functions (IEEE_CEC_2020) of different types and complexity levels. The search performance of the algorithm was analyzed using the Friedman and Wilcoxon statistical test methods. According to the analysis results, the WChOA-FDB variants consistently outperform the base algorithm across all tested dimensions, with Friedman score improvements ranging from 17.3% (Case-6) to 25.2% (Case-4), indicating that the FDB methodology provides significant optimization enhancement regardless of problem complexity. Additionally, experimental evaluations conducted on color image segmentation tasks demonstrate the effectiveness of the proposed algorithm in achieving accurate and efficient segmentation results. The WChOA-FDB method demonstrates significant improvements in Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), and Feature Similarity Index (FSIM) metrics with average enhancements of 0.121348 dB, 0.012688, and 0.003676, respectively, across different threshold levels (m = 2 to 12), objective functions, and termination criteria.
ISSN:2073-8994
DOI:10.3390/sym17071066
Fuente:Engineering Database