期刊文献+
任意字段
题名或关键词
题名
关键词
文摘
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息

EM算法的改进及其在行为识别中的应用 被引量：3

Application of Improved EM Algorithm in Recognition of Human Actions
下载PDF
导出
摘要 EM算法是求解GMM参数的传统算法,当样本数据规模比较大、GMM高斯成分数量比较高时,EM算法需要很长的时间才能收敛。提出了一种改进的EM算法,通过设置适当的参数,利用改进后的EM算法求解GMM参数,相比原EM算法在运行速度上有了很大的提高;进一步,结合GMM超向量以及SVM分类器,将改进后的EM算法应用到对KTH人体行为数据库的识别中,相比原EM算法识别准确率只受到了很小的影响。 A long time is needed to estimate GMM parameters using the general EM algorithm when the training data is large and the number of Gauss components is big. An improved EM algorithm is proposed in this paper. Compared to the original algorithm the operating speed is significantly improved using the improved EM algorithm to estimate GMM parameters if the appropriate parameters are given. The improved EM algorithm is applied on the KTH human actions database combined with GMM supervector and SVM, and the recognition accuracy is affected only a little compared to the original EM algorithm.
作者 赵桂儒 李卫东 刘典婷 吴敏 崔满丰
机构地区 中国地震台网中心 迈阿密大学
出处 《电视技术》 北大核心 2014年第13期196-199,共4页 Video Engineering
基金 国家留学基金资助项目(201204190040) 地震行业专项(201208010)
关键词 高斯混合分布 EM算法 行为识别 GMM EM recognition of human action
分类号 TP391 [自动化与计算机技术—计算机应用技术]
  • 相关文献

参考文献13

  • 1CAMPBELL W M, STURIM D E, REYNOLDS D A. Support vector machines using GMM supervectors for speaker verification[J]. IEEE Signal Processing Letters, 2006,13 (5) : 308-311.
  • 2INOUE N, SHINODA K. A fast MAP adaptation technique for GMM-supervector-based video semantic indexing systems[C]//Proc. 19th ACM International Conference on Multimedia. [S.I.]: ACM Press,2011 : 1357-1360.
  • 3LIU D, SHYU M L, ZHAO G. Spatial-temporal motion information integration for action detection and recognition in non-static baek- ground[C]//Proc. 2013 IEEE 14th International Conference on Infor- mation Reuse and Integration (IRI). [S.1.]: IEEE Press, 2013: 626-633.
  • 4陈晓,张尤赛,邹维辰.一种基于GMM的图像语义标注方法[J].电视技术,2012,36(23):35-38. 被引量：1
  • 5钟金琴,辜丽川,檀结庆,李莹莹.基于分裂EM算法的GMM参数估计[J].计算机工程与应用,2012,48(34):28-32. 被引量：13
  • 6SCHULDT C, LAPTEV I,CAPUTO B. Recognizing human actions: a local SVM approach[C]//Proc. 7th International Conference on Pat- tern Recognition. [S.1.] : IEEE Press, 2004: 32-36.
  • 7LAPTEV I. On space-time interest points[J]. International Journal of Computer Vision, 2005,64 (2/3) : 107-123.
  • 8LOWED G. Distinctive image features from scale-invariant key- points[J]. International Journal of Computer Vision, 2004, 60 (2) : 91-110.
  • 9BOUMAN C A,SHAPIRO M,COOK G W,et al. Cluster: an unsuper- vised algorithm for modeling Gaussian mixtures[EB/OL]. [2014-Ol-lO].http://dynamo.ecn.purdue.edu/-bouman/software/ clus- ter.
  • 10INOUE N, SAITO T, SHINODA K, et al. High-level feature extrac- tion using sift GMMs and audio models[C]//Proc. 2010 20th Inter- national Conference on Pattern Recognition (ICPR). [S.I.]: IEEE Press, 2010 : 3220-3223.

二级参考文献20

  • 1王平波,蔡志明,刘旺锁.混合高斯概率密度模型参数的期望最大化估计[J].声学技术,2007,26(3):498-502. 被引量：21
  • 2LAVRENKO V. A model for learning the semantics of pictures [ C ]// Proceedings of International Conference on Advances in Neutral Informa- tion Processing Systems. [ S. 1. ] :IEEE Press,2004:553-560.
  • 3DUYGULU P, BARNARD K, DE FREITAS J F G, et al. Object recog- nition as machine translation:Learning a lexicon for a fixed image vocabu- lary[ C]//Proc. the 7th European Conference on Computer Vision-Part IV. Copenhagen:[ s. n. ] ,2002:97-112.
  • 4ESCALANTE H J, MONTES M, SUCAR E. Word co-occurrence and Markov random fields for improving automatic image annotation [ C ]// Proc. British Machine Vision Conference on machine vision and related areas. [ S. 1. ] :IEEE Press,2007:75-78.
  • 5BLEI D M,NG A Y,JORDAN M I. Latent dirichlet allocation[J]. Jour- nal of Machine Learning Research, 2003,3 (5) :993-1022.
  • 6PILLAR S, LINDSAY B G. Ahemative EM methods for nonparametric finite mixture models[J]. Biometrika, 2001,88(2) :535-550.
  • 7Data for "Object recognition as machine translation" ( ECCV 02 ) [ EB/ OL]. [ 2012-04-01 ]. http://kobus. Ca/research/data/eccv_2OO2/in- dex. html.
  • 8Acharya U R, Sree S V.Application of recurrence quanti- fication analysis for the automated identification of epi- leptic EEG signals[J].Intemational Journal of Neural Sys- tems(IJNS) ,2011,21 (3) : 199-211.
  • 9Fabian T.Mixture of Gaussians exploiting histograms of oriented gradients for background subtraction[C]//LNCS 6454: ISVC' 10 Proceedings of the 6th International Con- ference on Advances in Visual Computing, 2010 , part II: 716-725.
  • 10Zhang Y,Michael S.Effective online unsupervised adap- tation of Gaussian Mixture Models and its application to speech classification[J].Pattern Recognition Letters, 2008,29 (6) : 735-744.

共引文献12

同被引文献28

  • 1岳佳,王士同.高斯混合模型聚类中EM算法及初始化的研究[J].微计算机信息,2006,22(11X):244-246. 被引量：51
  • 2王维彬,钟润添.一种基于贪心EM算法学习GMM的聚类算法[J].计算机仿真,2007,24(2):65-68. 被引量：15
  • 3Yang M S,Lai C Y.A robust EM clustering algorithm for Gaussian mixture models[J].Patter Recognition,2012,45(10):3950-3961.
  • 4Dempster A P,Laird N M.Max likelihood from incomplete data via the EM algorithm[J].Journal of the Royal Statistical Society Series B,1997,39(4):1-38.
  • 5Reddy K,Chiang H D,Rajaratnam B.TRUST-TECH-based expectation maximization for learning finite mixture models[J].Computational Statistic&Data Analysis,2013,41(4):561-575.
  • 6Figueiredo M A T,Jain A K.Unsupervised learning of finite mixture models[J].IEEE Transactions on Pattern Analysis,2002,24(6):381-396.
  • 7Lo K,Gottardo R.Flexible mixture modeling via the multivariate distribution with the Box-Cox transformation:an alternative to the skew-t distribution[J].Statistics and Computing,2012,18(5):33-52.
  • 8Caudill S B.A partially adaptive estimator for the censored regression model based on a mixture of normal distributions[J].Statistical Methods and Applications,2012,21(8):121-137.
  • 9Li J Q,Barron A.Mixture density estimation[J].Advances in Neural Information Processing Systems,2000,15(1):279-285.
  • 10何勇,刘青宝.基于动态网格的数据流聚类分析[J].计算机应用研究,2008,25(11):3281-3284. 被引量：6

引证文献3

二级引证文献10

电视技术
2014年 第13期

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