Abstract
In this paper we propose a deep learning framework to identify anomalous events in surveillance videos. Anomalous events are those which do not adhere to normal behaviour. We propose to use two discriminatively trained Convolutional Neural Networks, to capture the spatial and temporal features of videos, the classification scores obtained from the two streams are later fused to assign one final score. Since our approach is scenario-based, this eliminates the need for adopting a particular definition of anomaly. We show that the Two Stream CNNs perfectly capture the intricacies involved in modelling a video data by demonstrating the framework on airport and mall surveillance datasets respectively. We achieve a final test accuracy of 99.1% for spatial stream and 91% for temporal stream for airport scenario and an accuracy of 94.7% for spatial and 90.1% for the temporal stream for the mall scenario. Our framework can be easily implemented in real-time and is capable of detecting anomaly in each frame fed by a live surveillance system.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Li, T., Chang, H., Wang, M., Ni, B., Hong, R., Yan, S.: Crowded scene analysis: a survey. arXiv:1502.01812v1 [cs.CV]. Last accessed 6 Feb 2015
Karpathy, A., Toderici, G., Shetty, S., Leung, T., Sukthankar, R., Fei-Fei, L.: Large-scale video classification with convolutional neural networks. In: CVPR (2014)
Simonyan, K., Zisserman, A.: Two-stream convolutional networks for action recognition in videos. In: Proceedings of the 27th International Conference on Neural Information Processing Systems, vol. 1 (2014)
Srivastava, N., Mansimov, E., Salakhudinov, R.: Unsupervised learning of video representations using LSTMs. arXiv:1502.04681v3 [cs.LG]. Last accessed 4 Jan 2016
Medel, J.R., Savakis, A.: Anomaly detection in video using predictive convolutional long short-term memory networks. arXiv:1612.00390; Cong, Y., Yuan, J., Liu, J.: Sparse reconstruction cost for abnormal event detection. In: CVPR, pp. 3449–3456 (2011)
Cong, Y., Yuan, J., Liu, J.: Sparse reconstruction cost for abnormal event detection. In: CVPR, pp. 3449–3456 (2011)
Mahadevan, V., Li, W., Bhalodia, V., Vasconcelos, N.: Anomaly detection in crowded scenes. In: Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (2010)
Feichtenhofer, C., Pinz, A., Zisserman, A.: Convolutional two-stream network fusion for video action recognition. arXiv:1604.06573v2 [cs.CV]. Last accessed 26 Sep 2016
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the inception architecture for computer vision. arXiv:1512.00567 [cs.CV]
Donahue, J., Jia, Y., Vinyals, O., Hoffman, J., Zhang, N., Tzeng, E., Darrell, T.: DeCAF: a deep convolutional activation feature for generic visual recognition. arXiv:1310.1531v1 [cs.CV]. Last Accessed 6 Oct 2013
Farneback, G.: Two-frame motion estimation based on polynomial expansion. In: Proceedings of the 13th Scandinavian Conference on Image Analysis (2003); Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.Y., Berg, A.C.: SSD: single shot multibox detector. arXiv:1512.02325 [cs.CV]
Lin, K., Chen, S.-C., Chen, C.-S., Lin, D.-T., Hung, Y.-P.: Abandoned object detection via temporal consistency modelling and back-tracing verification for visual surveillance. IEEE Trans. Inf. Forensic Secur. (TIFS) (2015)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Jamadandi, A., Kotturshettar, S., Mudenagudi, U. (2020). Two Stream Convolutional Neural Networks for Anomaly Detection in Surveillance Videos. In: Elçi, A., Sa, P., Modi, C., Olague, G., Sahoo, M., Bakshi, S. (eds) Smart Computing Paradigms: New Progresses and Challenges. Advances in Intelligent Systems and Computing, vol 766. Springer, Singapore. https://doi.org/10.1007/978-981-13-9683-0_5
Download citation
DOI: https://doi.org/10.1007/978-981-13-9683-0_5
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-13-9682-3
Online ISBN: 978-981-13-9683-0
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)