首页 | 本学科首页   官方微博 | 高级检索  
     

基于多分类SVM的图像情感分类研究
引用本文:刘全中,王吉军. 基于多分类SVM的图像情感分类研究[J]. 大连轻工业学院学报, 2008, 27(4): 329-332
作者姓名:刘全中  王吉军
作者单位:大连大学信息工程学院,辽宁大连116622
基金项目:辽宁省教育厅科研项目(20060040); 大连市科技基金项目(2004-166-1)
摘    要:依据基于二叉树思想的多分类支持向量机算法,进行了天气图像的情感分类方法的研究。通过定性分析和计算对比,确定了图像情感分类的可行算法和合理参数选取方法。结果表明,完全二叉决策树SVM算法对于天气图像的情感识别具有较理想的效果。

关 键 词:图像情感  决策树  多分类  支持向量机

Image emotional classification based on multiclass SVM
LIU Quan-zhong,WANG Ji-junInformation Engineering College,Dalian University,Dalian,,China. Image emotional classification based on multiclass SVM[J]. Journal of Dalian Institute of Light Industry, 2008, 27(4): 329-332
Authors:LIU Quan-zhong  WANG Ji-junInformation Engineering College  Dalian University  Dalian    China
Affiliation:LIU Quan-zhong,WANG Ji-junInformation Engineering College,Dalian University,Dalian,116622,China
Abstract:The emotional classified methods of images were investigated in the paper according the idea of binary tree based multi-class SVM.The feasible algorithm of image emotional classification and reasonable parameter select method were proposed after qualitative analysis and calculation.Results showed that fully binary decision tree algorithm has good emotional classification rate for image emotional recognition.
Keywords:image emotion  decision-tree  multi-classification  SVM  
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号