期刊文献+

视频语义分类特征选择算法 被引量:5

Algorithm of Feature Selection for Semantic Video Classifier
下载PDF
导出
摘要 提出一种将连续特征数字量化后进行特征选择的算法(ABFSA)。利用样本集中的先验信息选择出特征值域中最具类间区分意义的区域,将其作为完整量化区间。采用向后式的启发搜索策略,搜索合并后能使贝叶斯分类错误率降低的相邻量化区间。合并搜索得到的两相邻量化区间,量化的级数降低一阶。重复搜索和合并过程,直至贝叶斯分类错误率不再降低为止。所有特征搜索、合并完成后,总的特征量化阶数得到大幅降低。用UCI仿真数据集及真实视频数据进行实验,对比结果表明该算法能有效选取视频语义概念分类的重要特征,其综合性能较优。 Classifier performance could be improved by selecting the most important features from high dimension feature set. A method of classification based on approximated Bayesian error feature selection algorithm (ABFSA) was proposed.In order to get the distinctive value partition of feature, the neighborhood of mean for each class was preserved. Then, the residual value field of feature was divided. Backward sequential feature selection, a heuristic search strategy was used in the merging step of quantity partition. At the same time, quantity complexity was remarkably reduced. The results of experiments comparing to other feature selection methods indicate the algorithm can effectively select features.
出处 《系统仿真学报》 EI CAS CSCD 北大核心 2006年第5期1143-1146,共4页 Journal of System Simulation
基金 国家自然科学基金项目资助(60273035) 江苏省科技攻关项目(BE2003064)
关键词 特征选择 特征量化 贝叶斯分类错误率 视频语义概念 feature selection feature quantity Bayesian classification error rate semantic video concept
  • 相关文献

参考文献8

  • 1Jensen R,Shen Q.Semantics-Preserving Dimensionality Reduction:Rough and Fuzzy-Rough-Based Approaches[J].IEEE Transactions on Knowledge and Data Engineering(S1041-4347),2004,16(12):1457-1471.
  • 2Rasheed Z,Sheikh Y,Shah M.On the Use of Computable Features for Film Classification[J].IEEE Transactions on Circuits and Systems for Video Technology(S1051-8215),2005,15(1):52-64.
  • 3Liu H,Setiono R.Feature selection via discretization[J].IEEE Transactions on Knowledge and Data Engineering (S1041-4347),1997,9(4):642-645.
  • 4Mittal A,Cheong L-F.Addressing the Problems of Bayesian Network Classification of Video Using High-dimensional Features[J].IEEE Transactions on Knowledge and Data Engineering (S1041-4347),2004,16(2):230-244.
  • 5JTC1/SC29/WG11 I.I.Coding of Moving Pictures and Audio,"Overview of the MPEG-7 Standard" Int'l Organization for Standariation.Oct.2000[S].
  • 6Lynch R S J,P K W.Bayesian Classification and Feature Reduction Using Uniform Dirichlet Priors[J].IEEE Transactions on Systems,Man and Cybernetics(S1083-4419),Part B,2003,33(3):448-464.
  • 7Ji H,Bang S Y.Feature Selection for Multi-class Classification Using Pairwise Class Discriminatory Measure and Covering Concept[J].Electronics Letters (S0013-5194),2000,36(6):524-525.
  • 8C.Blake C M.UCI Machine Learning Repository[DB/OL].(2005)[2005].http://www.ics.uci.edu/~mlearn/MLRepository.html.

同被引文献57

引证文献5

二级引证文献22

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部