Study on image feature selection: a genetic algorithm approach
Huanzhou Huang, Yating Wu, Yung‐Kuan Chan, Chaoheng Lin · 2010
This study was mainly about genetic algorithms of feature selection. The features adopted by this paper include CCM and DBPSP for the relationship between color and texture, CHKM for the color information of an image. The genetic algorithm of this study is implemented by MatLab program. The genetic algorithm optimization and searching technology adopted mechanics of genes and natural selection, and the algorithm implementation steps are: population initialization, fitness functions, selection, crossover, mutation, iteration and evolution. Feature selections used here are Sequential Forward Selection (SFS), Sequential Backward Selection (SBS), and the genetic algorithm-based feature selection used in this essay respectively. This study was analyzed and compared the result of the experiment respectively. The experiment was carried out for comparing the image retrieval accuracy, feature selection and computing time of image retrieval.