Enhancing Color Image Segmentation through a Hybrid PSO-GWO Algorithm with Graph-Based Refinement

Image segmentation is a critical task in computer vision with applications in medical imaging, remote sensing, surveillance, and object recognition. Traditional segmentation techniques, including thresholding and clustering, often fail on noisy or complex color images with low contrast or irregular textures. To address these challenges, this paper proposes a hybrid framework integrating Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), and graph-cut refinement within a multi-objective optimization setting for image segmentation. Our proposed PSO-GWO hybrid algorithm aims to combine the rapid convergence of PSO with the adaptive exploration of GWO, enhancing solution diversity and robustness. This optimizer is embedded into a Strength Pareto Evolutionary Algorithm 2 (SPEA-2) framework to simultaneously optimize intra-region homogeneity, edge strength, and spatial connectivity. Subsequently, graph-cut refinement and edge detection techniques enforce spatial continuity and sharpen segmentation boundaries. The proposed method is evaluated on the benchmark BSDS300 dataset and compared with existing approaches, measuring probabilistic rand index (PRI). The experimental results report notable PRI values, evidencing significant improvements in segmentation quality and boundary precision. These outcomes achieve performance comparable to, and in certain aspects surpassing, state-of-the-art methods, underscoring the effectiveness of the proposed framework.

• Supervision