Scholarlink

<p>This paper presents a novel hybrid metaheuristic algorithm called Whale-Ant Lion Optimization (WALO) for solving the multilevel image thresholding problem. The proposed approach combines the exploration capabilities of Ant Lion Optimizer (ALO) with the exploitation strengths of Whale Optimization Algorithm (WOA) to determine optimal threshold values based on Kapur’s entropy. We further enhance the algorithm’s performance by incorporating opposition-based learning strategies, resulting in two variants: WALO with opposition-based learning (WALO-OBL) and WALO with random opposition-based learning (WALO-ROBL). Extensive experiments are conducted on standard test images with threshold levels ranging from 1 to 10. The performance of the proposed algorithms is evaluated using multiple metrics including standard deviation of fitness values, mean square error, peak signal-to-noise ratio, structural similarity index, and feature similarity index. Experimental results demonstrate that the proposed WALO variants significantly outperform existing state-of-the-art algorithms in terms of both solution quality and convergence speed. Specifically, WALO achieved up to <b>84%</b> reduction in fitness standard deviation compared to ALO, while WALO-OBL demonstrated improvements of approximately <b>15–20%</b> in FSIM scores at higher thresholds. The WALO-ROBL variant showed superior performance with MSE reductions of up to <b>32%</b> compared to conventional methods, particularly excelling in finding optimal thresholds while maintaining computational efficiency across different threshold levels with convergence typically stabilizing within 30–40 iterations.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A hybrid whale–ant lion optimization algorithm with random opposition-based strategy for multilevel image thresholding

  • 期刊论文

  • 发表日期：

    2025年12月19日

查看全文
  • S. Abhay Sastha,
  • Manish Mahato,
  • Sarada Mohapatra,
  • Subhabrata Rath,
  • Prabhujit Mohapatra

摘要

This paper presents a novel hybrid metaheuristic algorithm called Whale-Ant Lion Optimization (WALO) for solving the multilevel image thresholding problem. The proposed approach combines the exploration capabilities of Ant Lion Optimizer (ALO) with the exploitation strengths of Whale Optimization Algorithm (WOA) to determine optimal threshold values based on Kapur’s entropy. We further enhance the algorithm’s performance by incorporating opposition-based learning strategies, resulting in two variants: WALO with opposition-based learning (WALO-OBL) and WALO with random opposition-based learning (WALO-ROBL). Extensive experiments are conducted on standard test images with threshold levels ranging from 1 to 10. The performance of the proposed algorithms is evaluated using multiple metrics including standard deviation of fitness values, mean square error, peak signal-to-noise ratio, structural similarity index, and feature similarity index. Experimental results demonstrate that the proposed WALO variants significantly outperform existing state-of-the-art algorithms in terms of both solution quality and convergence speed. Specifically, WALO achieved up to 84% reduction in fitness standard deviation compared to ALO, while WALO-OBL demonstrated improvements of approximately 15–20% in FSIM scores at higher thresholds. The WALO-ROBL variant showed superior performance with MSE reductions of up to 32% compared to conventional methods, particularly excelling in finding optimal thresholds while maintaining computational efficiency across different threshold levels with convergence typically stabilizing within 30–40 iterations.