一种运动数据检索的相关反馈算法

doi:10.3969/j.issn.0372-2112.2016.04.016

PDF(979 KB)

电子学报 ›› 2016, Vol. 44 ›› Issue (4) : 868-872. DOI: 10.3969/j.issn.0372-2112.2016.04.016

学术论文

一种运动数据检索的相关反馈算法

陈松乐^1,2, 孙正兴¹, 张岩¹, 李骞¹

作者信息 +

A Relevance Feedback Algorithm for Motion Data Retrieval

CHEN Song-le^1,2, SUN Zheng-xing¹, ZHANG Yan¹, LI Qian¹

Author information +

文章历史 +

摘要

本文提出了一种基于RankBoost的运动数据检索相关反馈算法.该算法具有以下二个方面的特点:首先,以KNN-DTW作为RankBoost集成学习的弱排序器,在适应变长多变量时间序列(Variable-Length Multivariate Time Series,VLMTS)数据的同时,利用RankBoost的集成性与高效性解决相关反馈实时性要求与VLMTS数据计算复杂度高的矛盾;其次,以本文提出的最小化排序经验损失和泛化损失风险作为RankBoost集成学习目标,有效地克服了相关反馈小样本学习环境下的过拟合问题.在CMU动作库上的实验结果验证了该方法的有效性.

Abstract

A relevance feedback algorithm based on RankBoost for content-based motion data retrieval (CBMR) is presented and has two characteristics.First,KNN-DTW is employed as the weak ranker for RankBoost ensemble learning.While adapting to variable-length multivariate time series (VLMTS) data,by taking the advantage of the ensemble and efficiency of RankBoost,it can resolve the conflict between the real-time requirement of relevance feedback and the high computational complexity of VLMTS data.Second,minimizing ranking experience loss and generalization loss risk proposed in this paper are used as the learning objective for RankBoost ensemble learning,which can effectively solve the over-fitting problem caused by small-sample training in relevance feedback.Experimental results on CMU action library verify the effectiveness of the proposed algorithm.

导出引用

陈松乐, 孙正兴, 张岩, 李骞. 一种运动数据检索的相关反馈算法[J]. 电子学报, 2016, 44(4): 868-872. https://doi.org/10.3969/j.issn.0372-2112.2016.04.016

CHEN Song-le, SUN Zheng-xing, ZHANG Yan, LI Qian. A Relevance Feedback Algorithm for Motion Data Retrieval[J]. Acta Electronica Sinica, 2016, 44(4): 868-872. https://doi.org/10.3969/j.issn.0372-2112.2016.04.016

中图分类号： TP391.41

参考文献

[1] LIU F,ZHUANG Y T,Wu F,et al.3D motion retrieval with motion index tree[J].Computer Vision and Image Understanding,2003,92(2-3):265-284.
[2] CHEN S L,SUN Z X,LI Y,et al.Partial similarity human motion retrieval based on relative geometry features[A].Proceedings of 4th International Conference on Digital Home[C].Washington:IEEE Computer Society,2012.298-303.
[3] TANG J K T,LEUNG H.Retrieval of logically relevant 3D human motions by adaptive feature selection with graded relevance feedback[J].Pattern Recognition Letters,2012,33(4):420-430.
[4] HUANG T S,ZHOU X S.Image retrieval with relevance feedback:from heuristic weight adjustment to optimal learning methods[A].Proceedings of 2001 International Conference on Image Processing[C].Washington:IEEE Computer Society,2001.2-5.
[5] ZHOU X S,HUANG T S.Relevance feedback in image retrieval:a comprehensive review[J].Multimedia Systems,2003,8(6):536-544.
[6] WU K,YAP K H.Fuzzy SVM for content-based image retrieval[J].IEEE Computational Intelligence Magazine,2006,1(2):10-16.
[7] YOON H,YANG K,SHAHABI C.Feature subset selection and feature ranking for multivariate time series[J].IEEE Transactions on Knowledge and Data Engineering,2005,17(9):1186-1198.
[8] HUANG W,GAO Y,CHAN K L.A review of region-based image retrieval[J].Journal of Signal Processing Systems for Signal Image and Video Technology,2010,59(2):143-161.
[9] LEI H S,SUN B Y.A study on the dynamic time warping in kernel machines[A].Proceedings of the 2007 Third International IEEE Conference on Signal-Image Technologies and Internet-Based System[C].Washington:IEEE Computer Society,2007.839-845.
[10] KOHAVI R,SOMMERFIELD D.Feature subset selection using the wrapper method:overfitting and dynamic search space topology[A].Proceedings of 1st International Conference on Knowledge Discovery and Data Mining[C].California:Amer Assn for Artificial,1995.192-197.
[11] HUANG S H,WU Q J,LAI S H.Improved AdaBoost-based image retrieval with relevance feedback via paired feature learning[J].Multimedia Systems,2006,12(1):14-26.
[12] JIANG W,ER G,DAI Q H,GU J W.Similarity-based online feature selection in content-based image retrieval[J].IEEE Transactions on Image Processing,2006,15(3):702-712.
[13] ZHOU X S,GARG A,HUANG T S.Nonlinear variants of biased discriminants for interactive image retrieval[J].Vision,Image and Signal Processing,IEE Proceedings,2005,152(6):927-936.
[14] FREUND Y,IYER R,SCHAPIRE R E,et al.An efficient boosting algorithm for combining preferences[J].Journal of Machine Learning Research,2003,4:933-969.
[15] CMU.Motion Capture Database[DB/OL].http://mocap.cs.cmu.edu/,2003.
[16] MULLER M,RODER T,CLAUSEN M.Efficient content-based retrieval of motion capture data[J].ACM Transactions on Graphics,2005,24(3):677-685.
[17] SAKOE H,CHIBA S.Dynamic programming algorithm optimization for spoken word recognition[J].IEEE Transactions on Acoustics,Speech and Signal Processing,1978,26(1):43-49.
[18] LIU T Y.Learning to rank for information retrieval[J].Foundations and Trends in Information Retrieval,2009,3(3):225-331.
[19] CRUCIANU M,FERECATU M,BOUJEMAA N.Relevance feedback for image retrieval:a short survey[R].Basel:DELOS2 European Network of Excellence(FP6),2004.

基金

国家高科技发展计划 (No.2007AA01Z334); 国家自然科学基金 (No.61272219,No.61100110,No.61321491); 教育部新世纪优秀人才资助计划 (No.NCET-04-0460); 江苏省科技计划 (No.BE2010072,No.BE2011058,No.BY2012190,No.BY2013072-04); 计算机软件新技术国家重点实验室创新基金重点项目 (No.ZZKT2013A12)