Abstract
In this paper, we propose two novel semantic event detection models, i.e., Two-dependence Bayesian Network (2d-BN) and Conditional Random Fields (CRFs). 2d-BN is a simplified Bayesian Network classifier which can characterize the feature relationships well and be trained more efficiently than traditional complex Bayesian Networks. CRFs are undirected probabilistic graphical models which offer several particular advantages including the abilities to relax strong independence assumptions in the state transition and avoid a fundamental limitation of directed probability graphical models. Based on multi-modality fusion and mid-level keywords representation, we use a three-level framework to detect semantic events. The first level extracts audiovisual features, the mid-level detects semantic keywords, and the high-level infers events using 2d-BN and CRFs models. Compared with state of the art, extensive experimental results demonstrate the effectiveness of the proposed two models.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Cestnik: Estimating probabilities: A crucial task in machine learning. In: Proc. 9th European Conf. on Artificial Intelligence, pp. 147–149 (1990)
Snoek, C.G., Worring, M., Smeulders, A.: Early versus late fusion in sematic video analysis. In: ACM Multimedia Conference, pp. 399–402 (2005)
Ekin, A.M.T., Mehrotr, R.: Automatic soccer video analysis and summarization. IEEE Trans. on Image processing 12(7), 796–807 (2003)
FlexCRFs: Flexible Conditional Random Fields, http://www.jaist.ac.jp/~hieuxuan/flexcrfs/flexcrfs.html
Bai, H.L., Hu, W., Wang, T., Tong, X.F., Zhang, Y.M.: A Novel Sports Video Logo Detector Based on Motion Analysis. In: International Conference on Neural Information Processing (ICONIP) (2006)
Intel Open Source Probabilistic Network Library (OpenPNL), http://www.intel.com/research/mrl/pnl
Lafferty, J., McCallum, A., Pereira, F.: Conditional random fields: probabilistic models for segmenting and labeling sequence data. In: Proc. of ICML, pp. 282–289 (2001)
Wang, J., Xu, C., Chng, E., Wan, K., Tian, Q.: Automatic replay generation for soccer video broadcasting. In: ACM Multimedia Conference (2004)
Duan, L., Xu, M., Chua, T.-S., Tian, Q., Xu, C.: A mid-level representation framework for semantic sports video analysis. In: ACM Multimedia Conference (2003)
Xie, L., Chang, S.-F., Divakaran, A., Sun, H.: Structure analysis of soccer video with hidden markov models. Proc. ICASSP 4, 4096–4099 (2002)
LIBSVM: A Library for Support Vector Machines, http://www.csie.ntu.edu.tw/~cjlin/libsvm/
Luo, M., Ma, Y., Zhang, H.J.: Pyramidwise structuring for soccer highlight extraction. In: ICICS-PCM, pp. 1–5 (2003)
Xu, M., Maddage, N., Xu, C., Kankanhalli, M., Tian, Q.: Creating audio keywords for event detection in soccer video. In: IEEE ICME 2003, vol. 2, pp. 281–284 (2003)
Friedman, N., Geiger, D., Goldszmidt, M.: Bayesian network classifiers. Machine Learning 29(2), 131–163 (1997)
Sha,, Pereira., F.: Shallow parsing with conditional random fields. In: Proc. of HLT/NAACL (2003)
Li, X.K., F.M.: A hidden Markov model framework for traffic event detection using video features. In: IEEE Proc. of ICIP 2004, pp. 2901–2904 (2004)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2006 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Wang, T., Li, J., Hu, W., Tong, X., Zhang, Y., Dulong, C. (2006). Event Detection Models Using 2d-BN and CRFs. In: Cham, TJ., Cai, J., Dorai, C., Rajan, D., Chua, TS., Chia, LT. (eds) Advances in Multimedia Modeling. MMM 2007. Lecture Notes in Computer Science, vol 4352. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-69429-8_9
Download citation
DOI: https://doi.org/10.1007/978-3-540-69429-8_9
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-69428-1
Online ISBN: 978-3-540-69429-8
eBook Packages: Computer ScienceComputer Science (R0)