Retinal Vessel Segmentation Combined Two-Dimensional Entropy Method and Double Populations Genetic Algorithm

In order to overcome the disadvantages such as finite sampling space and local optimal of genetic algorithm, the main objective of this paper is to combine double populations genetic algorithm and two-dimensional maximum entropy threshold method for retinal vessels segmentation. The proposed method is able to segment retinal vessels image accurately and keep connectivity and smoothness of vessels through the numerical experiments. Numerical results show that the combined algorithm has faster convergence speed, higher calculation accuracy, stronger noise resistance and better performance in reserving pathological information compared with other algorithms.