Evaluation of genetic algorithm for interactive evolutionary face image beautifying system

The performance of a genetic algorithm (GA) with a new way of crossover proposed by authors before is evaluated quantitatively to verify its advantage. This new algorithm was proposed for interactive evolutionary face image beautifying system implemented on a small mobile device. Since the display size is small, GA must work effectively for a very small population size. The new crossover adopts internally and externally dividing and was highly evaluated subjectively by questionnaires. In order to clarify the advantage of the new algorithm, the performance is evaluated by the mean square error between the output face image and the desired face image obtained by modifying the original face image with a graphical tool. The result shows that the performance of the new algorithm is higher than the conventional methods for GA.