Optimized Histogram Equalization for Image Enhancement

Abstract

In this project, Image Enhancement has been achieved by performing Histogram Equalization that uses optimization algorithms to optimize parameters.Histogram equalization is a spatial domain image enhancement technique, which effectively enhances the contrast of an image.However, while it takes care of contrast enhancement,it does not consider the abrupt changes in the image brightness due to which image brightness is not preserved.Hence,in this project a modified histogram equalization technique using optimization algorithm has been proposed, which takes care of contrast enhancement while ensuring brightness preservation.The idea used here is to first ,section the data image histogram into two, utilizing otsu's limit .Then an arrangement of streamlined measuring requirements are formed and connected on both the sub-images. Then, the sub-images are evened out freely and their union creates the contrast enhanced , brightness preserved output image .Here we have used three Optimization Algorithms for finding the optimal constraints . First , Genetic Algorithm(GA) has been used , to optimise the constraints .Second , Particle Swarm Optimization (PSO) has been used and third ,a Hybrid PSO Optimization Algorithm has been used for the same .Then the results produced by the above algorithms are compared to find out which one outperforms the other , by comparing various parameters like Discrete Entropy , Mean , Number of Generations.