Image Segmentation Based On Maximum Fuzzy Entropy And Genetic Algorithm

Image segmentation is one of basis problem of digital image processing and machine vision, and it is also an important step for detecting and identifying objects. The main difficulties lie in the great variability of images and the presence of noises.Recently, many researchers have introduced fuzzy set theory and genetic algorithm to image segmentation, which can get better results than traditional algorithms. An improved adaptive genetic algorithm in image segmentation is proposed to improve image division performance and division efficiency based on the study of fuzzy threshloding methods.Firstly, because the simple genetic algorithm is easily premature, an improved adaptive genetic algorithm is proposed. This method adopts a new variable to evaluate the concentration degree of population fitness. According to the concentration degree, the crossover probability and mutation probability is adaptively changed, which could improve the convergence of the genetic algorithm.Then, a thresholding method for image segmentation is presented, based on two-dimensional maximum fuzzy entropy and genetic algorithm. Utilizing two-dimensional histogram, the method defines a membership function that is fitter for image characteristics, and then gives the description of image's fuzzy entropy. The procedure for finding the optimal combination of fuzzy parameters is implemented by simple genetic algorithm and improved adaptive genetic algorithm. Finally, the optimal threshold is determined by maximizing the fuzzy entropy. The experimental results indicated that the proposed method gave better performance and higher calculation speed, and the ability of resisting noise is improved. For multi-target image segmentation, a three-level thresholding method is presented. The method defines different membership functions for dark part, gray part and bright part of the image, and then gives the definition of fuzzy entropy. An improved adaptive genetic algorithm is proposed for the optimization of fuzzy parameters. Finally, the optimal thresholds can be determined by maximizing the fuzzy entropy. The experimental results demonstrated that the proposed method could segment the image effectively and fast.