Abstract
In this paper, a novel probabilistic topic model is proposed for mining activities from complex video surveillance scenes. In order to handle the temporal nature of the video data, we devise a dynamical causal topic model (DCTM) that can detect the latent topics and causal interactions between them. The model is based on the assumption that all temporal relationships between latent topics at neighboring time steps follow a noisy-OR distribution. And the parameter of the noisy-OR distribution is estimated by a data driven approach based on the idea of nonparametric Granger causality statistic. Furthermore, for convergence analysis during model learning process, the Kullback-Leibler between the prior and the posterior distributions is calculated. At last, using the causality matrix learned by DCTM, the total causal influence of each topic is measured. We evaluate the proposed model through experimentations on several challenging datasets and demonstrate that our model can identify the high influence activity in crowded scenes.
Similar content being viewed by others
References
Blei DM, Lafferty JD (2006) Dynamic topic models[C] Proceedings of the 23rd international conference on machine learning. ACM, pp 113–120
Blei DM, Ng AY, Jordan MI (2003) Latent dirichlet allocation[J]. J Mach Learn Res 3:993–1022
Chua FCT, Oentaryo RJ, Lim EP (2015) Using linear dynamical topic model for inferring temporal social correlation in latent Space[J]. Computer Science
Emonet R, Varadarajan J, Odobez J (2011) Extracting and locating temporal motifs in video scenes using a hierarchical non parametric Bayesian model[C]. IEEE Computer Vision and Pattern Recognition 3233–3240
Fan Y, Yang H, Zheng S et al (2013) Video sensor-based complex scene analysis with Granger causality[J]. Sensors 13(10):13685–13707
Faruquie TA, Kalra PK, Banerjee S (2009) Time based activity inference using latent dirichlet allocation[C]. BMVC 1–10
Gu B, Sheng VS (2016) A robust regularization path algorithm for v-support vector classification[J]. IEEE Transactions on Neural Networks and Learning Systems 28(5):1241-1248
Gu B, Sun X, Sheng VS (2016) Structural minimax probability machine. IEEE Transactions on Neural Networks and Learning Systems. doi:10.1109/TNNLS.2016.2544779
Gu B, Sheng VS, Tay KY et al (2015) Incremental support vector learning for ordinal regression[J]. IEEE Transactions on Neural Networks and Learning Systems 26(7):1403–1416
Gu B, Sheng VS, Wang Z et al (2015) Incremental learning for v-support vector regression[J]. Neural Netw 67:140–150
Hofmann T (2001) Unsupervised learning by probabilistic latent semantic analysis[J]. Mach Learn 42(1-2):177–196
Hospedales T, Gong S, Xiang T (2009) Markov clustering topic model for mining behaviour in video[C] IEEE international conference on computer vision, IEEE, pp 1165–1172
Kinoshita A, Takasu A, Adachi J (2015) Real-time traffic incident detection using a probabilistic topic model[J]. Inf Syst 54(C):169–188
Kuettel D, Breitenstein MD, Gool LV et al (2010) What’s going on? Discovering spatio-temporal dependencies in dynamic scenes[C]. IEEE Computer Vision and Pattern Recognition 1951–1958
Kular D, Ribeiro E (2015) Analyzing activities in videos using latent dirichlet allocation and granger causality[C] International symposium on visual computing. Springer International Publishing, pp 647–656
Li J, Gong S, Xiang T (2008) Global behaviour inference using probabilistic latent semantic analysis[C]. BMVC 3231:3232
Li Y, Lu H, Li J et al (2016) Underwater image de-scattering and classification by deep neural network[J]. Comput Electr Eng 54:68–77
Lu H, Li Y, Nakashima S et al (2016) Single image dehazing through improved atmospheric light estimation[J]. Multimedia Tools and Applications 75(24):17081–17096
Mccaffery JP, Maida AS (2013) Toward a causal topic model for video scene analysis[C] International joint conference on neural networks, pp 1–8
Nedungadi A, Rangarajan G, Jain N, Ding M (2008) Analyzing multiple spike trains with nonparametric granger causality. J Comput Neurosci 27:55–64
Qian S, Zhang T, Xu C et al (2016) Multi-modal event topic model for social event analysis[J]. IEEE Trans Multimedia 18(2):233–246
Song L, Jiang F, Shi Z et al (2011) Understanding dynamic scenes by hierarchical motion pattern mining[C] IEEE international conference on multimedia and expo (ICME), 2011. IEEE, pp 1–6
Teh YW, Jordan MI, Beal MJ, Blei DM (2006) Hierarchical dirichlet processes. J Am Stat Assoc 101(476)
Varadarajan J, Emonet R, Odobez JM (2013) A sequential topic model for mining recurrent activities from long term video Logs[J]. Int J Comput Vis 103 (1):100–126
Wang X, Ma X, Grimson WEL (2009) Unsupervised activity perception in crowded and complicated scenes using hierarchical bayesian models[J]. IEEE Trans Pattern Anal Mach Intell 31(3):539–555
Xue J, Eguchi K (2016) Sequential correspondence hierarchical dirichlet processes for video data analysis[C]. ACM ACM 229–233
Zach C, Pock T, Bischof H (2007) A duality based approach for realtime TV-l, 1 optical Flow[J]. Lect Notes Comput Sci 4713(5):214–223
Zhou L, Hu RQ, Qian Y et al (2013) Energy-spectrum efficiency tradeoff for video streaming over mobile ad hoc networks[J]. IEEE J Sel Areas Commun 31 (5):981–991
Acknowledgements
The authors would like to thank the associated editor and all the anonymous reviewers for their valuable comments and suggestions. This work was partly supported by Natural Science Foundation of Jiangsu Province(Grant No. BK20160908), and the Key Lab of Broadband Wireless Communication and Sensor Network Technology(Grant No.NYKL2015012), and the National Natural Science Foundation of China (Grant No. 61401228, 61501253) and the Basic Research Program of Jiangsu Province(Natural Science Foundation)(Grant No. BK20151506), and the China Postdoctoral Science Foundation(Grant No. 2015M581841), and the Postdoctoral Science Foundation of Jiangsu Province (Grant No. 1501019A), and the Priority Academic Program Development of Jiangsu Higer Education Institutions(PAPD), and the Jiangsu Collaborative Innovation Center on Atmospheric Environment and Equipment Technology(CICAEET), and the Nanjing University of Information Science and Technology Research Foundation for Talented Scholars (Grant No. 2015r014).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Fan, Y., Zhou, Q., Yue, W. et al. A dynamic causal topic model for mining activities from complex videos. Multimed Tools Appl 77, 10669–10684 (2018). https://doi.org/10.1007/s11042-017-4760-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-017-4760-4