基于改进的RS-GA图像特征选择方法  

Image feature selection method based on improved RS-GA

作　　者：张杰慧[1] 何中市[1] 黄丽琼[1] 

机构地区：[1]重庆大学计算机学院,重庆400044

出　　处：《计算机应用》2006年第10期2372-2374,共3页journal of Computer Applications

基　　金：国家自然科学基金资助项目(60173060)

摘　　要：针对目前图像识别中原始特征数量大、不相关特征多以及冗余等现象,提出了一种图像特征选择方法。将遗传算法(GA)与粗集(RS)思想有机结合进行特征选择,引入粗集中相关属性依赖度的定义,设计了适应度函数和遗传算子,以提高算法时间效率和获得最佳搜索结果,并将该特征选择方法应用于图像,实验表明,基于改进的RS-GA图像特征选择方法达到了较好的效果,并具有较高的算法效率。With regard to the problem that original feature in image classification is mass and redundancy, a new image feature selection method was presented. This method combines the Rough Set（RS） theory with Genetic Algorithm（GA） properly to select feature. To improve the efficiency of this algorithm and get the optimal searching result, definition of relative attribute dependency of rough set theory was introduced, and fitness function and genetic operators were designed. Then, this proposed method was applied to image feature selection. Experimental results show that it has better performance and higher algorithm efficiency.

关 键 词：粗集理论 属性约简 图像特征选择 遗传算法 相关属性依赖度 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]