The week's pick
Reseach Article
Revolutionizing Video Surveillance: AI-Powered Anomaly Detection
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 186 - Number 74 |
| Year of Publication: 2025 |
| Authors: Nishant Deheriya, Devendra Kumar Bajpai, P.K. Sharma |
10.5120/ijca2025924606
|
Nishant Deheriya, Devendra Kumar Bajpai, P.K. Sharma . Revolutionizing Video Surveillance: AI-Powered Anomaly Detection. International Journal of Computer Applications. 186, 74 ( Mar 2025), 37-41. DOI=10.5120/ijca2025924606
Abstract
The widespread deployment of surveillance cameras in public spaces such as airports, roadways, and financial institutions has led to the generation of massive volumes of video data. However, a significant portion of this footage is reviewed only after an incident occurs, making manual monitoring both inefficient and error-prone. Automated anomaly detection in video surveillance has thus emerged as a critical area of research, aiming to enhance security by identifying abnormal activities in real-time. This study proposes a novel deep learning-based framework for anomaly detection in video streams by integrating an Inflated 3D Convolution Network (I3D-ResNet50) with deep Multiple Instance Learning (MIL). The model treats video sequences as collections of instances, where normal and anomalous segments are classified as negative and positive instances, respectively. Each video snippet is individually assessed using a fully connected neural network (NN) to compute an anomaly score, enabling precise identification of abnormal activities. To enhance feature extraction and generalization, the I3D-ResNet50 model is applied after performing 10-crop augmentations on the UCF-101 dataset, which comprises approximately 50 GB of video data spanning 15 types of anomalous events such as fighting, theft, and abuse, alongside normal activities. Extensive experiments demonstrate the effectiveness of our approach, achieving an Area Under the Curve (AUC) score of 83.85% within just 10,000 iterations, significantly outperforming conventional anomaly detection techniques. The proposed model exhibits robustness in detecting subtle and complex anomalies in dynamic environments, making it well-suited for real-time surveillance applications. By reducing the dependency on manual monitoring and improving anomaly detection accuracy, this system has the potential to enhance public safety and security across various domains, including transportation hubs, banking institutions, and smart cities.
References
- M. T. Subha Devi, M. Dhanalakshmi, S. A, S. ML and L. N, "Anomaly Detection in Video Surveillance," 2024 IEEE 9th International Conference for Convergence in Technology (I2CT), Pune, India, 2024, pp. 1-3, doi: 10.1109/I2CT61223.2024.10543949.
- Vajda, D.L., Do, T.V., Bérczes, T. et al. Machine learning-based real-time anomaly detection using data pre-processing in the telemetry of server farms. Sci Rep 14, 23288 (2024). https://doi.org/10.1038/s41598-024-72982-z.
- Proceedings of the IEEE International Conference on Computer Vision, 2017, v. 2017-October, page number. 341-349.DOI: http://dx.doi.org/10.1109/ICCV.2017.45.
- Edmund Fosu Agyemang.Anomaly detection using unsupervised machine learning algorithms: A simulation study.Scientific African Volume 26, December 2024, ScienceDirect.
- Hamza Karim, Keval Doshi, Yasin Yilmaz;Real-Time Weakly Supervised Video Anomaly Detection. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024, pp. 6848-6856.
- Wang, T., Liu, Z., & Liu, L. (2024). Investigating a three-dimensional convolution recognition model for acoustic emission signal analysis during uniaxial compression failure of coal. Geomatics, Natural Hazards and Risk, 15(1).https://doi.org/10.1080/19475705.2024.2322483.
- Ullah W, Ullah A, Haq IU, Muhammad K, Sajjad M, Baik SW (2021) Cnn features with bi-directional LSTM for real-time anomaly detection in surveillance networks. Multimedia Tools and Applications 80(11):16979–16995.
- Zhong Y, Chen X, Jiang J, Ren F. A cascade reconstruction model with generalization ability evaluation for anomaly detection in videos. Pattern Recogn 2022. 122:108336.
- Tian Y, Pang G, Chen Y, Singh R, Verjans JW, Carneiro G. Weakly-supervised video anomaly detection with robust temporal feature magnitude learning. In: Proceedings of the IEEE/CVF International Conference on Computer Vision 2021, pp. 4975–4986.
- Lu C, Shi J, Jia J. Abnormal event detection at 150 fps in matlab. In: Proceedings of the IEEE International Conference on Computer Vision 2013,pp. 2720–2727.
- Sultani W, Chen C, Shah M. Real-world anomaly detection in surveillance videos. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 6479–6488.
- Zisserman A, Carreira J, Simonyan K, Kay W, Zhang B, Hillier C, Vijayanarasimhan S, Viola F, Green T, Back T, et al (2017) The kinetics human action video dataset.
- Carreira J, Zisserman A (2017) Quo vadis, action recognition? a new model and the kinetics dataset. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 6299–6308.
- Zhao B, Fei-Fei L, Xing EP (2011) Online detection of unusual events in videos via dynamic sparse coding. In: CVPR 2011, pp. 3313–3320. IEEE.
- Dubey S, Boragule A, Jeon M (2019) 3d resnet with ranking loss function for abnormal activity detection in videos. In: 2019 International Conference on Control, Automation and Information Sciences (ICCAIS), pp. 1–6. IEEE.
Index Terms
Keywords