The Evaluation of Nature-Inspired Optimization Techniques for Contrast Enhancement in Images: A Novel Software Tool

This study is rooted in the direct correlation between the performance of multivariate techniques and the selection of parameters. The complexity and time-consuming nature of parameter selection, due to the need for exhaustive testing of all available parameters for optimal results, is acknowledged. To mitigate this issue, a novel software tool, integrating nine nature-inspired optimization methods (Differential Evolution, Artificial Bee Colony, Particle Swarm, Cat Swarm, Dragonfly, Black Hole, Bacterial Foraging, Genetic Algorithms, and Simulated Annealing), is proposed. These methods are employed in histogram stretching, a parameter-dependent contrast enhancement technique, with multiplication, addition, and root extraction operations as the target parameters for optimization. In addition to this, histogram equalization, a parameter-independent contrast enhancement technique, is included for the purpose of comparative performance analysis. The software tool, publicly available, provides four performance metrics namely, Mean Square Error, Peak Signal-to-noise Ratio, Structural Similarity Index, and processing times. A rigorous evaluation using the widely recognized Tampere Image dataset indicates that Differential Evolution emerged as the most efficient technique, scoring highest for Structural Similarity Index (0.948) and second best for Mean Square Error (278.05) and Peak Signal to Noise Ratio (26.962). Furthermore, Particle Swarm Optimization demonstrated the fastest time complexity, requiring merely 0.6 sec per image for parameter definition. Notably, it was observed that while histogram equalization tends to degrade original images, the adaptive nature of optimized histogram stretching remains preserved, thereby leaving the image quality unaffected. Such findings highlight the efficacy of the proposed software tool in the optimization and evaluation of contrast enhancement techniques.