期刊文献+

一种基于多互补分类器的自动视频语义标注方法 被引量:1

An Automatic Video Annotation Method Based on Multiple Complementary Classifiers
下载PDF
导出
摘要 在现有互训练(Co-Training)算法的基础上,提出了一种基于多个互补型分类器的半监督学习(Semi-Supervised Learn-ing)方法,并将其应用到自动视频语义标注框架中.该方法通过构建基于特征互补和模型互补的多个分类器对未标注样本中的隐含信息加以利用,并结合视频序列中概念分布的时间相关性和局部聚集性等特性提高了分类的准确性,相对于有监督学习方法提高了约7%左右. A novel semi-supervised learning method based on multiple complementary classifiers is proposed according to the analysis of the existing co-training algorithms, which is applied to the automatic video semantic annotation. By constructing the classifiers with feature complementarity and model complementarity,the hidden information in unlabeled dataset is effectively exploited. Moreover,the local consistency and temporal relationship of video sequences is further applied to improve the final annotation accuracy, the improvement of annotation accuracy is about 7%.
出处 《小型微型计算机系统》 CSCD 北大核心 2007年第11期2085-2089,共5页 Journal of Chinese Computer Systems
基金 国家自然科学基金项目(603333020)资助.
关键词 视频语义标注 半监督学习 互训练算法 video semantic annotation semi-supervised learning co-training
  • 相关文献

参考文献17

  • 1Wu J,Hua X.-S,Zhang B,et al.An online-optimized incremental learning framework for video semantic classification[C].ACM Mukimedia,320-323,New York,October 10-16,2004.
  • 2Xie L,et al.Structure analysis of soccer video with hidden markov models[C].ICASSP,Orlando,Florida,May,2002,4,13-17.
  • 3Fan J,et al.Semantic video classification by integrating flexible mixture model with adaptive EM algorithm[C].ACM Multimedia MIR,Berkeley,CA,November 2-8,2003,9-16.
  • 4Dempster A P,Laird N M,Rubin D B.Maximum likelihood from incomplete data via the EM algorithm[J].J.Royal Statistical Soc.,Series B,1977,39(1):1-38.
  • 5Joachims T.Transductive inference for text classification using support vector machines[C].Proc.16th Int'l Conf.on Machine Learning,pp.200-209,Bled,Slovenia,June 27-30,1999.
  • 6Blum,Mitchell T.Combining labeled and unlabeled data with cotraining[C].Proceedings of the Workshop on Computational Learning Theory,92-100,1998.
  • 7Goldman S,Y Z.Enhancing supervised learning with unlabeled data[C].In Proc.of the 17th Int'l Conf.on Machine Learning,Stanford University,June 29-July 2,2000,327-334,.
  • 8Dasgupta S,Littman M,McAllester D.PAC generalization bounds for co-training[A].Advances in Neural Information Processing Systems[M].Cambridge,Mass:MIT Press,2002,14,375-382.
  • 9Rosenberg C,Heberg M,Schneiderman H.Semi-supervised selftraining of object detection models[C].7th IEEE Workshop on Applications of Computer Vision,Breckenridge,CO,Jan,2005.
  • 10Goldman S,Zhou Y.Enhancing supervised learning with unlabeled data[C].Proc.17th Int'l Conf.Machine Learning,Stanford University,June 29-July 2,2000,327-334.

同被引文献10

引证文献1

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