Author(s)

DengBin Xu¹, PeiChen Wu¹, LiNing Yuan^2*

Affiliation(s)

¹ School of Information Network Security, People's Public Security University of China, Beijing 100038, China.

² School of Public Security Big Data Modern Industry, Guangxi Police College, Nanning 530028, Guangxi, China.

Corresponding Author

LiNing Yuan

ABSTRACT

The performance of current attention based feature fusion methods depends on the correlation between features. After feature fusion, due to the inter domain differences of different features, the spatiotemporal perception ability is insufficient, and effectively fusing two cross domain features still faces challenges. A video anomaly detection method based on compression inflation attention feature fusion is proposed to address the issues of insufficient cross domain expression ability of RGB features and optical flow features, as well as weak spatiotemporal perception ability of fused features. The use of Squeeze and Inflation Networks (SENet) to construct a fusion mechanism for RGB and optical flow features can enhance the expression ability of fused features while reducing the number of network parameters and improving the performance of anomaly detection algorithms. In the global spatiotemporal awareness stage, the ConvLSTM (Long Short Term Memory Convolutional Network) is used to achieve global spatiotemporal awareness, while balancing computational complexity and detection performance. We achieved a recognition performance of 93.72% on the UCSDPed2 dataset and also performed well on the CUHKAvenue and LAD2000 datasets, verifying the effectiveness of the method.

KEYWORDS

Computer vision; Abnormal behaviour detection; Feature fusion; Attentional mechanism; Multi-branch convolution

CITE THIS PAPER

DengBin Xu, PeiChen Wu, LiNing Yuan. Video anomalous behaviour detection based on compressed-inflated attention module. Journal of Computer Science and Electrical Engineering. 2024, 6(2): 22-28. DOI: 10.61784/jcsee3005.

REFERENCES

[1] Jiancong Wang. Deep learning based multi-module joint video anomaly detection research. Guilin University of Electronic Science and Technology, 2023.

[2] Liu Yang, Yang Dingkang, Wang Yan, et al. Generalised video anomaly event detection: systematic taxonomy and comparison of deep models. arXiv preprint arXiv: 2302.05087, 2023.

[3] Chang Yunpeng, Tu Zhigang, Xie Wei, et al. Video anomaly detection with spatio-temporal dissociation. Pattern Recognition, 2022, 122: 108213.

[4] Tudor Ionescu r, Smeureanu Sorina, Alexe Bogdan, et al. Unmasking the abnormal events in video. Proceedings of the 2017 IEEE international conference on computer vision. Piscataway: IEEE, 2017: 28952903.

[5] Sultani Waqas, Chen Chen, Shah Mubarak. Real-world anomaly detection in surveillance videos. Proceedings of the 2018 IEEE conference on computer vision and pattern recognition. Piscataway: IEEE, 2018: 6479-6488.

[6] Wan B, Fang Y, Xia X, et al. Weakly supervised video anomaly detection via centre-guided discriminative learning//2020 IEEE international conference on multimedia and expo (ICME). IEEE, 2020: 1-6.

[7] Liu Wen, Luo Weixin, Li Zhengxin, et al. Margin Learning Embedded Prediction for Video Anomaly Detection with A Few Anomalies. Proceedings of the 2019 International Joint Conferences on Artificial Intelligence. Freiburg: IJCAI, 2019: 3023-3030.

[8] Wan Boyang, Jiang Wenhui, Fang Yuming, et al. Anomaly detection in video sequences: a benchmark and computational model. IET Image Processing, 2021, 15(14): 3454-3465.

[9] Zhou Jiapeng. Research on Unsupervised Learning-based Domain-adaptive Semantic Segmentation Methods. China University of Mining and Technology, 2023.

[10] Zou Wei, Zhang Dong, Lee Dahjye. A new multi-feature fusion based convolutional neural network for facial expression recognition. Applied Intelligence, 2022, 52(3): 2918-2929.

[11]  Cheng Xianggui, Liu Zhao, Guo Fang. Video Abnormal Event Detection Combining Dual-stream I3D and Attention Mechanism. Information and Computer (Theoretical Edition), 2022, 34(24): 65-68.

[12] Dai Yimian, Gieseke Fabian, Oehmcke Stefan, et al. Attentional feature fusion. Proceedings of the 2021 IEEE/CVF winter conference on applications of computer vision. Piscataway: IEEE, 2021: 3560-3569.

[13]  Huang Shaonian, WEN Peiran, QUAN Qi, et al. Lightweight video anomaly detection based on multi-branch aggregation frame prediction. Journal of Graphics, 2023, 44(6): 1173.

[14]  Le Viettuan, Kim Yongguk. Attention-based residual autoencoder for video anomaly detection. Applied Intelligence, 2023, 53: 3240-3254.

[15] Ye Wenbing, Zhan Shihua. Anomaly Detection Method of Network Traffic Based on MHA-BiLSTM. Modern Information Technology, 2024, 8(02): 65-69.

[16] Liu Wei, Rabinovich Andrew, C.Berg Alexander. Parsenet: Looking wider to see better. arXiv preprint arXiv:1506.04579. 2015.

[17] Shi Xingjia, Chen Zhourong, Wang Hao, et al. Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting. Proceedings of the 2015 Conference on Neural Information Processing Systems. San Diego: NeurIPS, 2015.

[18] Szegedy Christian, Liu Wei, Jia Yangqing, et al. Going deeper with convolutions. Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition, Boston. Washington: IEEE Computer Society, 2015: 1-9.

[19] Yan Shanwu, Xiao Hongbing, Wang Yu, et al. Video anomaly detection by fusing pedestrian spatiotemporal information. Journal of Graphics, 2023, 44(1): 95.

[20]  Roka Sanjay, Diwakar Manoj: a deep convolutional encoder-decoder architecture for abnormality detection in video surveillance. Cluster Computing, 2024: 1-16.

[21]  Li Kunchang, Wang Yali, He Yinan, et al. Uniformerv2: Unlocking the potential of image vits for video understanding. Proceedings of the 203 IEEE/CVF International Conference on Computer Vision. Piscataway: IEEE, 2023: 1632-1643.

[22] Hu Jie, Shen Li, Sun Gang. Squeeze-and-excitation networks. Proceedings of the 2018 IEEE conference on computer vision and pattern recognition. Piscataway: IEEE, 2018: 7132- 7141.

[23]  Lv Hui, Yue Zhongqi, Sun Qianru, et al. Unbiased multiple instance learning for weakly supervised video anomaly detection. Proceedings of the 2023 IEEE/CVF conference on computer vision and pattern recognition. Piscataway: IEEE, 2023: 8022-8031.

[24]  Park Hyunjong, Noh Jongyoun, Ham Bumsub. Learning memory-guided normality for anomaly detection. Proceedings of the 2020 IEEE/CVF conference on computer vision and pattern recognition. Piscataway: IEEE, 2020: 14372-14381.