中北大学学报（自然科学版）2017，Vol.38Issue(2)：196-201,6.DOI:10.3969/j.issn.1673-3193.2017.02.018
基于最优特征加权的图像分类算法
Image Classification Algorithm Based on Optimal Feature Weighting
摘要
Abstract
This article focuses on the problem of feature selection and classifier construction in the image classification process.An image classification algorithm based on optimal feature weighting was proposed.Firstly,collect the color and texture feature of images.Secondly,calculate the weights of each feature using chaotic particle swarm optimization algorithm,where the weights represents contributions of correspond feature to the classification model is determined.Finally,the weighted vector sample set is studied by the relevant vector machine,and then construct the classifier to achieve the image classification.Experiment results show that the proposed algorithm not only can improve the accuracy of image classification,but also reduce the classification time,which means that it has a certain practical value compared with the popular image classification model.关键词
图像分类/特征选择/混沌粒子群算法/相关向量机Key words
image classification/features selection/particles swarm optimization algorithm/relevance vector machine分类
信息技术与安全科学引用本文复制引用
王玉晶..基于最优特征加权的图像分类算法[J].中北大学学报（自然科学版）,2017,38(2):196-201,6.基金项目
四川省教育厅重点科研资助项目(15ZA0339) （15ZA0339）
阿坝师范学院校级科研基金资助项目(ASC15-20) （ASC15-20）
