Image Feature Selection Based on Genetic Algorithm

摘要

Image feature selection is formulated as an optimization problem. When traditional genetic algorithm is used for selecting image feature, it may bring problems of local convergence or precocious puberty because of using a fixed probability of crossover operator and mutation operator. First, the paper gave a brief introduction to image feature selection regarding the purposes, tasks, and commonly used algorithms. Then, the paper improved genetic operator of genetic algorithm, and parallel computing was used to genetic algorithm, to enhance the performance of image feature selection. Finally, experiments show that the improved genetic algorithm is convergent and effective, and applied to image feature selection is successful.