Crowd Anomaly Detection Using Deep Learning

Abstract— This paper introduces an image-based crowd counting method within the Audiovisual Crowd framework, designed to detect and warn against mass crowd stampedes or trampling caused by dangerous overcrowding. The approach utilizes deep convolution neural networks (CNNs) to analyze both visual and audio data, extracting meaningful features to identify anomalies in crowd behavior. The model is trained using CNN architectures and evaluated with Mean Absolute Error (MAE) and Mean Squared Error (MSE) metrics to ensure precise crowd estimation. Experimental results show that while the video-only approach effectively captures spatial information, it is more susceptible to challenges such as low-light and noisy environments. In contrast, the combined audiovisual model achieves enhanced robustness and accuracy, reaching an overall accuracy of 94% in detecting critical crowd anomalies.

Crowd Counting, Density Estimation, Deep Learning, CNN

1. Asgary, A. (2021). World crowd disaster map. [Google Scholar] [Crossref]

2. https://yorku.maps.arcgis.com/apps/webappviewer/index.html?id=e7c52856187642e19bd227865393432c [Google Scholar] [Crossref]

3. V. Joshi, N. M. Patel. Supervised Deep Learning Approaches for Anomaly Detection and Recognition in Crowd Scenes. Electronic Letters on Computer Vision and Image Analysis. 2025;24(1):31-50. [Google Scholar] [Crossref]

4. Claudio Feliciani, Alessandro Corbetta, Milad Haghani, Katsuhiro Nishinari, Trends in crowd accidents based on an analysis of press reports, Safety Science, Volume 164, 2023, 106174, ISSN 0925-7535, [Google Scholar] [Crossref]

5. S. Roy. A Comprehensive Survey on Crowd Anomaly Detection Using Deep Learning. Preprint. 2024; June:1-20. doi:10.13140/RG.2.2.24390.48961. [Google Scholar] [Crossref]

6. M. Ali, M. Al-Berry, Z. Taha. Comparative Study for Anomaly Detection in Crowded Scenes. Ain Shams University Reports. 2021;1(1):1-15. [Google Scholar] [Crossref]

7. Mehmood. Efficient Anomaly Detection in Crowd Videos Using Pre-Trained 2D Convolutional Neural Networks. IEEE Access. 2021; 9:14000-14012. doi:10.1109/ACCESS.2021.3118009. [Google Scholar] [Crossref]

8. W. Yaqub, M. Mohanty, B. Suleiman. Privacy-Preserving Online Proctoring Using Image-Hashing Anomaly Detection. University of Sydney Research Reports. 2022;1(1):1-12. [Google Scholar] [Crossref]

9. K. V. Joshi, N. M. Patel. A CNN Based Approach for Crowd Anomaly Detection. Gujarat Technological University Research Reports. 2022;1(1):1-10. [Google Scholar] [Crossref]

10. F. de-la-Calle-Silos, I. Gonzalez-Díaz, F. Díaz-de-María. Mid-Level Feature Set for Specific Event and Anomaly Detection in Crowded Scenes. Universidad Carlos III de Madrid Research Papers. 2013;1(1):1-20. [Google Scholar] [Crossref]

11. H. Guo, X. Wu, N. Li, R. Fu, G. Liang, W. Feng. Anomaly Detection and Localization in Crowded Scenes Using Short-term Trajectories. Conference/Technical Report. 2013;1(1):1-12. [Google Scholar] [Crossref]

12. Q. Zhou, J. Zhang, L. Che, H. Shan, and J. Z. Wang, ‘‘Crowd counting with limited labeling through submodular frame selection,’’ IEEE Trans. Intell. Transp. Syst., vol. 20, no. 5, pp. 1728–1738, May 2019. [Google Scholar] [Crossref]

13. J. Fu, H. Yang, P. Liu, and Y. Hu, ‘‘A CNN-RNN neural network join long short-term memory for crowd counting and density estimation,’’ in Proc. IEEE Int. Conf. Adv. Manuf. (ICAM), Nov. 2018, pp. 471–474, doi:10.1109/AMCON.2018.8614939. [Google Scholar] [Crossref]

14. R. Hu, S. Chang, H. Wang, J. He, and Q. Huang, ‘‘Efficient multispike learning for spiking neural networks using probability-modulated tim-ing method,’’ IEEE Trans. Neural Netw. Learn. Syst., vol. 30, no. 7, pp. 1984–1997, Jul. 2019. [Google Scholar] [Crossref]

15. R. Hu, S. Zhou, Z. R. Tang, S. Chang, Q. Huang, Y. Liu, W. Han, and E. Q. Wu, ‘‘DMMAN: A two-stage audio–visual fusion framework for sound separation and event localization,’’ Neural Netw., vol. 133, pp. 229–239, Jan. 2021. [Google Scholar] [Crossref]

16. T. Afouras, J. S. Chung, A. Senior, O. Vinyals, and A. Zisserman, ‘‘Deep audio-visual speech recognition,’’ IEEE Trans. Pattern Anal. Mach. Intell., early access, Dec. 21, 2019, doi: 10.1109/TPAMI.2018.2889052. [Google Scholar] [Crossref]

17. D. Gebru, S. Ba, X. Li, and R. Horaud, ‘‘Audio-visual speaker diarization based on spatiotemporal Bayesian fusion,’’ IEEE Trans. Pattern Anal. Mach. Intell., vol. 40, no. 5, pp. 1086–1099, May 2018. [Google Scholar] [Crossref]

18. B. Wu and R. Nevatia, ‘‘Detection of multiple, partially occluded humans in a single image by Bayesian combination of edgelet part detectors,’’ in Proc. 10th IEEE Int. Conf. Comput. Vis. (ICCV), vol. 1, Oct. 2005, pp. 90–97. [Google Scholar] [Crossref]

19. R. Raja, P. C. Sharma, M. R. Mahmood, and D. K. Saini, ‘‘Analysis of anomaly detection in surveillance video: Recent trends and future vision, ’Multimedia Tools Appl., vol. 82, no. 8, pp. 12635–12651, Mar. 2023. [Google Scholar] [Crossref]

20. Chaudhary, C. Bhatt, and S. R. Parikh, ‘‘Video preprocessing and feature extraction for anomaly detection,’’ Int. J. Comput. Appl., vol. 167, no. 4, pp. 13–17, 2020. [Google Scholar] [Crossref]

21. Y. Benezeth, P. Jodoin, B. Emile, H. Laurent, and C. Rosenberger, ‘‘Review and evaluation of commonly-implemented background subtraction algo-rithms,’’ J. Electron. Imag., vol. 19, no. 1, Dec. 2008, Art. no. 011009. [Google Scholar] [Crossref]

22. Simonyan and A. Zisserman, ‘‘Very deep convolutional networks for large-scale image recognition,’’ 2014, arXiv:1409.1556. [Google Scholar] [Crossref]

23. S. Ren, K. He, R. Girshick, and J. Sun, ‘‘Faster R-CNN: Towards real-timeobject detection with region proposal networks,’’ in Proc. Adv. Neural Inf.Process. Syst., vol. 28, Dec. 2015, pp. 91–99. [Google Scholar] [Crossref]

24. K. Santhosh, D. P. Dogra, and P. P. Roy, ‘‘Anomaly detection in road traffic using visual surveillance: A survey,’’ ACM Comput. Surv., vol. 53, no. 6, pp. 1–26, Nov. 2020. [Google Scholar] [Crossref]

25. Bozcan and E. Kayacan, ‘‘Context-dependent anomaly detection for low altitude traffic surveillance,’’ in Proc. IEEE Int. Conf. Robot. Autom. (ICRA), May 2021, pp. 224–230. [Google Scholar] [Crossref]

26. S. Zhou, W. Shen, D. Zeng, and Z. Zhang, “Unusual event detection in crowded scenes by trajectory analysis,” in Acoustics, Speech and Signal Processing (ICASSP), 2015 IEEE International Conference on. IEEE,2015, pp. 1300–1304. [Google Scholar] [Crossref]

27. V. Reddy, C. Sanderson, and B. C. Lovell, “Improved anomaly detection in crowded scenes via cell-based analysis of foreground speed, size and texture,” in Computer Vision and Pattern Recognition Workshops (CVPRW), 2011 IEEE Computer Society Conference on. IEEE, 2011. [Google Scholar] [Crossref]

• Competency and Challenges of BTLED-ICT Students in 2D Animation: An Analytical Study
• Slope Stability Assessment: A Case Study of Embankments Along OMU-Aran-Ilorin Road, Nigeria
• Advancements in Precursors, Materials, Deposition Techniques for Thin Film Research in Electronic Devices: A Mini Review
• “Empowering Indian Women through Entrepreneurship: A Study on Kolkata”
• Impact of Mental Mathematics Proficiency on Job Performance Among Seconadry Schools Teachers in Emohua and Port Hacourt City