小差异化图像数据库中的特定特征挖掘方法设计  

Design of specific feature mining method in image database with small alienation

作　　者：刘平[1] 王晓[1] 刘春[2] 

机构地区：[1]河北科技大学图书馆,石家庄050018 [2]河北科技大学环境科学与工程学院,石家庄050018

出　　处：《沈阳工业大学学报》2017年第5期562-566,共5页Journal of Shenyang University of Technology

基　　金：河北省教育厅青年基金资助项目(SQ161142)

摘　　要：针对传统的特定特征关联挖掘方法存在挖掘效率低的问题,提出基于一种推荐模式的小差异化图像数据库中的特定特征数据挖掘方法.运用萤火虫优化支持向量机参数法,提取小差异化图像数据特定特征,解决相似关联问题,采用主成分分析方法对小差异化图像特征进行降维处理,利用Laplace预测分类方法对提取的小差异化图像特定特征进行推荐分类,之后对分类的特定特征按照推荐等级进行挖掘.结果表明,所提出的挖掘方法要优于传统挖掘方法,准确率及效率得到明显提高.Aiming at the problem that the traditional specific feature association mining method has low mining efficiency,a specific feature data mining method in the image database with small alienation based on a recommendation model was proposed. With the firefly parameter optimization method of support vector machine（ SVM）,the specific feature of image data with small alienation was extracted,and the similarity association problem was solved. The principal component analysis method was used to reduce the dimension of image feature association with small alienation,and the Laplace prediction classification method was adopted to recommend and classify the specific features of extracted image with small alienation. In addition,the specific feature after the classification was mined according to the recommended levels. The results show that the proposed mining method is superior to the traditional mining methods,and the accuracy rate and efficiency get obviously enhanced.

关 键 词：萤火虫算法 图像数据库 特定特征 挖掘方法 Laplace预测 支持向量机 主成分分析法 推荐分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]