An Approach based on Neural Networks, Multiple Instance Learning and PCA for Anomaly Detection in Video Surveillance

• Silas S. L. Pereira UECE / IFCE http://orcid.org/0000-0003-2535-3024
• J. E. Bessa Maia UECE https://orcid.org/0000-0002-4983-1724

DOI: https://doi.org/10.5753/semish.2021.15814

Abstract

Multiple Instance Learning (MIL) has become an attractive solution in video surveillance literature once it allows working with weakly supervised bases. This work proposes and evaluates a video anomaly detection approach based on binary classification with Multilayer Perceptron (MLP) neural networks and MIL paradigm. The experiments were performed from a set of I3D (Inflated 3D) features which corresponds to the benchmark dataset ShanghaiTech. We also explore the effect of compactness and essential data representation with the feature extraction technique Principal Component Analysis (PCA). The achieved results were competitive when compared with state of art.

Keywords: anomaly detection, video surveillance, Multiple Instance Learning, I3D features, Multilayer Perceptron

References

Ali, S. and Shah, M. (2008). Human action recognition in videos using kinematic features and multiple instance learning. IEEE transactions on pattern analysis and machine intelligence, 32(2):288–303.

Haykin, S. (2010). Neural networks and learning machines, 3/E. Pearson Education India.

Kamoona, A. M., Gosta, A. K., Bab-Hadiashar, A., and Hoseinnezhad, R. (2020). Multiple instance-based video anomaly detection using deep temporal encoding-decoding. arXiv preprint arXiv:2007.01548.

Kingma, D. P. and Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.

Kuhn, M., Johnson, K., et al. (2013). Applied predictive modeling, volume 26. Springer.

Li, T., Wang, Z., Liu, S., and Lin, W.-Y. (2021). Deep unsupervised anomaly detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 3636–3645.

Nayak, R., Pati, U. C., and Das, S. K. (2020). A comprehensive review on deep learning-based methods for video anomaly detection. Image and Vision Computing, page 104078.

Pawar, K. and Attar, V. (2019). Deep learning approaches for video-based anomalous activity detection. World Wide Web, 22(2):571–601.

Pereira, S. S. L. and Maia, J. E. (2021). Anomaly detection in surveillance video of natural environment. International Journal of Computer Applications, 183(1):1–7.

Rao, T. N., Girish, G., and Rajan, J. (2017). An improved contextual information based approach for anomaly detection via adaptive inference for surveillance application. In Proceedings of International Conference on Computer Vision and Image Processing, pages 133–147. Springer.

Ribeiro, M., Lazzaretti, A. E., and Lopes, H. S. (2018). A study of deep convolutional auto-encoders for anomaly detection in videos. Pattern Recognition Letters, 105:13–22.

Saha, B. N., Ray, N., and Zhang, H. (2009). Snake validation: A pca-based outlier detection method. IEEE Signal Processing Letters, 16(6):549–552.

Suarez, J. J. P. and Naval Jr, P. C. (2020). A survey on deep learning techniques for video anomaly detection. arXiv preprint arXiv:2009.14146.

Sultani, W., Chen, C., and Shah, M. (2018). Real-world anomaly detection in surveillance videos. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 6479–6488.

Wan, B., Fang, Y., Xia, X., and Mei, J. (2020). Weakly supervised video anomaly detection via center-guided discriminative learning. In 2020 IEEE International Conference on Multimedia and Expo (ICME), pages 1–6. IEEE.

Yun, K., Honorio, J., Chattopadhyay, D., Berg, T. L., and Samaras, D. (2012). Two-person interaction detection using body-pose features and multiple instance learning. In 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pages 28–35. IEEE.

Zhao, H., Lai, Z., Leung, H., and Zhang, X. (2020). Feature Learning and Understanding: Algorithms and Applications. Springer Nature.