Metaheuristic-Optimized Exposure-Based Sub-Image Histogram Equalization for Low-Contrast Color Image Enhancement

Abstract

Low-contrast images often suffer from poor visibility, loss of structural details, and reduced perceptual quality, affecting human interpretation and subsequent image analysis tasks. Histogram-based enhancement techniques improve contrast but may cause over-enhancement, brightness distortion, and unnatural appearance due to improper histogram partitioning. To address these limitations, an optimized low-contrast color image enhancement framework based on Exposure-based Sub-Image Histogram Equalization (ESIHE) is proposed. ESIHE is used as the core technique to preserve brightness and exposure characteristics, while optimization algorithms determine the histogram division threshold. Harmony Search Optimization (HSO), Grey Wolf Optimization (GWO), Whale Optimization Algorithm (WOA), Firefly Algorithm (FA), Grasshopper Optimization Algorithm (GOA), and hybrid Grey Wolf–Particle Swarm Optimization (HGWPSO) are employed to determine the optimal threshold. Experimental results show that optimized ESIHE with HSO outperforms non-optimized ESIHE, achieving an average PSNR of 19.44 and an SSIM of 0.844, compared with 16.50 and 0.5707, respectively. Entropy reaches 6.04, with improved edge preservation and lower NIQE scores.