计算机科学 ›› 2017, Vol. 44 ›› Issue (6): 283-289.doi: 10.11896/j.issn.1002-137X.2017.06.050

• 图形图像与模式识别 • 上一篇    下一篇

一种基于改进PLSA和案例推理的行为识别算法

涂宏斌,岳艳艳,周新建,罗锟   

  1. 华东交通大学轨道交通学院 南昌330013,华东交通大学国际学院 南昌330013,华东交通大学载运工具与装备教育部重点实验室 南昌330013,华东交通大学轨道交通学院 南昌330013
  • 出版日期:2018-11-13 发布日期:2018-11-13
  • 基金资助:
    本文受华东交通大学基金(2003414090),国家自然科学基金(51268015),江西省自然科学基金项目(20122BAB206027)资助

Novel Action Recognition via Improved PLSA and CBR

TU Hong-bin, YUE Yan-yan, ZHOU Xin-jian and LUO Kun   

  • Online:2018-11-13 Published:2018-11-13

摘要: 针对行为人发生的行为因遮挡或者自遮挡可能导致行为歧义性的问题,提出基于改进PLSA和案例推理算法的行为识别方法。该算法既可以克服传统PLSA算法中生成式模型对观察特征序列的独立性假设会导致过拟合的缺点,又可以消除由于遮挡等原因引起的歧义性带来的识别精度降低问题。实验表明该方法能有效地提高人体行为识别准确率。

关键词: 行为识别,歧义性行为,时空兴趣点,PLSA,案例推理

Abstract: In order to recognize the ambiguous action meaning in the same scene for occlusion,the improved PLSA and CBR algorithm are used to recognize the simple action according to the space-time interest point.The improved algorithm can not only overcome the shortcomings of the over fitting in the traditional PLSA algorithm owe to the generative model independence assumption of observation sequences,but also decrease recognition ambiguity.Experiments based on the proposed method are executed on the public databases such as KTH,Weizmann,UCF sports and the self-building database.The results show that the proposed method has the performance of validity and effectiveness.

Key words: Action recognition, Ambiguous behavior,Space-time interest point,PLSA(Probabilistic Latent Semantic Analysis),CBR(Case Based Reason)

[1] 张洪博.视频中的人体行为识别若干关键技术研究[D].厦门:厦门大学,2013.
[2] 何卫华.人体行为识别关键技术研究[D].重庆:重庆大学,2012.
[3] SHIAN-RU K,HOANG L U T,YONG-JIN L,et al.A Review on Video-Based Human Activity Recognition[J].Computers,2013,2:88-131.
[4] AGGARWAL K J,RYOO S M.Human activity analysis:A review[J].ACM Computing Surveys,2011,3(3):194-218.
[5] ZHANG J,GONG S.Action categorization by structural probabilistic latent semantic analysis[J].Computer Vision and Image Understanding,2010,4(8):857-864.
[6] ZHANG P,ZHANG Y N,THOMAS T,et al.Moving people tra-cking with detection by latent semantic analysis for visual surveillance applications[J].Multimedia Tools and Applications,2014,8(3):991-1021.
[7] GUO P,MIAO Z,SHEN Y,et al.Continuous human action re-cognition in real time[J].Multimedia Tools and Applications,2012,8(3):827-844.
[8] Li R T,ZHANG C F,ZHONG G H.A novel method of abnormal behaviors recognition[J].Applied informatics and Communication,PT I,Communications in Computer and Information Science,2011,4:79-86.
[9] ZHU X D,LIU Z J.Human Abnormal Behavior Recognition Based on Topic Hidden Markov Model[J].Computer Science,2012,9(3):251-255.(in Chinese) 朱旭东,刘志镜.基于主题隐马尔科夫模型的人体异常行为识别[J].计算机科学,2012,9(3):251-255.
[10] 谢飞.基于主题模型的人物行为识别[D].苏州:苏州大学,2014.
[11] LAPTEV I,CAPUTO B.Recognition of human actions.November 2011.http://www.nada.kth.se/cvap/actions.
[12] SCHULDT C,LAPTEV I,CAPITP B.Recognizing Human Actions:A Local SVM Approach[J].Proceedings of International Conference on Pattern Recognition,Icpr,2004,3:32-36.
[13] GORELICK L,BLANK M,SHECHTMAN E,et al.Weizmman actions asspace-time shapes.November 2011.http://www.wisdom.weizmann.ac.il/vision/SpaceTimeActions.html.
[14] ZELNIK-MANOR L,IRANI M.Weizmann event-based analysis of video.November 2011.http://www.wisdom.weizmann.ac.il/ vision/VideoAnalysis/Demos/EventDetec tion/EventDetection.html.
[15] University of Central Florida.UCF sports action dataset.February 2012.http://vision.eecs.ucf.edu/datasetsActions.html.
[16] ZHU Y,Chen W,Guo G.Evaluating spatiotemporal interestpoint features for depth-based action recognition[J].Image & Vision Computing,2014,2(8):453-464.
[17] MARN-JIMNEZ M J,Yeguas E,Blanca N P D L.Exploring STIP-based models for recognizing human interactions in TV videos[J].Pattern Recognition Letters,2013,4(15):1819-1828.
[18] XIANG T,GONG S,BREGONZIO M.Recognising action asclouds of space-time interest points[C]∥IEEE Conference on Computer Vision & Pattern Recognition.IEEE,2009:1948-1955.
[19] HARRIS C,STEPHENS M.A combined corner and edge detector[J].Proc of Fourth Alvey Vision Conference,1988,8(3):147-151.
[20] GOPARVAR-FARD M,HEYDARIAN A,Niebles J C.Vision-based action recognition of earthmoving equipment using spatio-temporal features and support vector machine classifiers[J].Advanced Engineering Informatics,2013,7(4):652-663.
[21] HOFMANN T.Probabilistic latent semantic indexing[J].International Acm Sigir Conference on Research and Development in Information Retrieval,1999,42(1):50-57.
[22] HOFMANN T.Learning the similarity of documents:an information-geometric approach to document retrieval and categorization[M]∥Leen T K,Dietterich T G,Tresp V.eds.,Advances in Neural Information Processing Systems,vol.12.The MIT Press,2000:914-920.
[23] HOFMANN T.Unsupervised learning by probabilistic latentsemantic analysis[J].Machine Learning,2001,2(1/2):177-196.
[24] NIEBLES J C,WANG H,Fei-fei L.1 Unsupervised Learning of Human Action Categories Using Spatial-Temporal Words[J].In Proc.BMVC,2008,9(3):299-318.
[25] FERNANDEZ-BELTRANR,Pla F.Incremental probabilistic La-tent Semantic Analysis for Video Retrieval[J].Image & Vision Computing,2015,8:1-12.
[26] KONG Y,ZHAN?ㄠ?????????戠牡????扡牰?ive learning codebook for action recognition[J].Pattern Recognition Letters,2011,2(8):1178-1186.
[27] GRIGORIOS T,ARISTIDIS L.The MinMax k-means clustering algorithm[J].Pattern Recognition,2014,7(7):2505-2516.
[28] DRIESEN J,VAN HAMME H.Modelling vocabulary acquisition,adaptation and generalization in infants using adaptive bayesian PLSA[J].Neurocomputing,2011,4(11):1874-1882.
[29] BASSOU N,KOTROPOULOS C.RPLSA:A novel updatingscheme for Probabilistic Latent Semantic Analysis[J].Compu-ter Speech & Language,2011,5(4):741-760.
[30] XIA L M,YANG B J,CAI N P.Recognition of Suspicious Activity Using Case-based Reasoning in Video Surveiuancep[J].Journal of Chinese Computer Systems,2014,5(8):1891-1896.(in Chinese) 夏利民,杨宝娟,蔡南平.监控视频中基于案例推理的人体可疑行为识别[J].小型微型计算机系统,2014,5(8):1891-1896.
[31] NGUYEN T V,SONG Z,YAN S.STAP:Spatial-Temporal Attention-Aware Pooling for Action Recognition[J].IEEE Transa-ctions on Circuits & Systems for Video Technology,2015,25(1):77-86.
[32] PAU AGUST V,TRAVER J,PLA F.Bag-of-words with aggregated temporal pair-wise word co-occurrence for human action recognition[J].Pattern Recognition Letters,2014,9(C):224-230.
[33] IOSIFIDIS A,TEFAS A,PITAS I.Discriminant Bag of Words based representation for human action recognition[J].Pattern Recognition Letters,2014,9:185-192.
[34] ALAZRAI R,MOWAFI Y,LEE C S G.Anatomical-plane-based representation for human-human interactions analysis[J].Pattern Recognition,2015,8(8):2346-2363.
[35] SAMANTA S,CHANDA B.Space-Time Facet Model for Human Activity Classification[J].IEEE Transactions on Multimedia,2014,6(6):1525-1535.
[36] FAHAD SHAHBAZ K,JOOST V D W,Muhammad A R,et al.Semantic pyramids for gender and action recognition[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2014,3(8):3633-3645.
[37] MARN-JIMNEZ M J,MUOZ-SALINAS R,YEGUAS-BO-LIVAR E,et al.Human interaction categorization by using audio-visual cues[J].Machine Vision & Applications,2014,5
No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!