Hybrid JADE-SPBO-Based Optimization for Improved Multilevel Image Segmentation
Hussam Nawwaf Fakhouri, Amjad A. Hudaib, Sandi N. Fakhouri, Mohannad S. Alkhalaileh, Ahmad Kamel AL Hwaitat · 2025
This paper introduces the Hybrid JADE-SPBO optimization algorithm for robust and efficient multilevel image segmentation, integrating JADE's adaptive mutation and crossover with SPBO's psychology-inspired learning for balanced exploration and exploitation. Tested on standard benchmark images, the proposed method achieves notable improvements over state-of-the-art metaheuristics, evidenced by higher PSNR, SSIM, and FSIM scores. At challenging segmentation thresholds, Hybrid JADE-SPBO attains PSNR values up to 22.50 compared to competing methods' 21.08, while similarly improving SSIM from around 0.82 to 0.89 and FSIM from 0.83 to 0.91. These results highlight the algorithm's ability to consistently detect precise thresholds and handle complex search spaces, offering a versatile tool for applications in image processing.