Sequentially Modified Gravitational Search Algorithm for Image Enhancement

Gravitational Search Algorithm (GSA) is based on the acceleration trend feature of objects with a mass towards each other and includes many interdependent parameters. The gravitational constant among these parameters influences the speeds and positions of the agents, meaning that the search capability depends on the largescale gravitational constant. The proposed new algorithm, which was obtained with the use of two operators at different times of the call and sequentially doing works, was named as Sequentially Modified ‎ Gravitational Search Algorithm (SMGSA). SMGSA is applied to 10 basic and 6 composite benchmark functions. Each function is run 30 times and the best, mean and median values are obtained. The achieved results are compared with the Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and GSA among the heuristic optimization algorithms. Between GSA and the operator for each function convergence speed, standard deviation and graphical comparisons are included. Beside this, by using the Wilcoxon signed rank test, the comparison of the averages of the data as two dependent groups of GSA and the new operators is performed. It is seen that the obtained results provided better results than the other methods. Additionally, in this study, SMGSA was applied to the transformation function among image enhancement techniques which are engineering applications. The success of this method has been increased by optimizing the parameters of the transformation function used. Effective improvement has been achieved in terms of both visual and information quality.