LIGHTWEIGHT 3D CNN WITH CHANNEL-SPATIAL ATTENTION FOR CHEATING BEHAVIOR RECOGNITION IN EXAM SURVEILLANCE VIDEOS
Keywords
DOI:
https://doi.org/10.34238/tnu-jst.15789Abstract
Automatic detection of cheating behaviors in examinations is essential for maintaining academic integrity, particularly in large-scale and online assessments. This study proposes a lightweight human action recognition model that is suitable for deployment on edge devices with limited computational resources. The proposed method extends the MobileNet architecture into a three-dimensional convolutional neural network by employing three-dimensional depthwise separable convolutions to efficiently extract spatial and temporal features from video data. In addition, a convolutional block attention module is incorporated to enhance the model's ability to focus on salient motion regions while suppressing irrelevant background information. Experimental results on an examination surveillance dataset containing six categories of cheating behaviors demonstrate that the proposed model requires only 3.2 million parameters and 1.8 billion floating-point operations for each 16-frame video clip, substantially reducing computational cost compared with the reference model. The model achieves an accuracy of 93.7% and a weighted F1-score of 0.932. These results confirm that the proposed model provides an effective balance between recognition accuracy and computational efficiency, making it well suited for real-time examination surveillance systems deployed on edge devices.Downloads
References
P. M. Newton and K. Essex, “How common is cheating in online exams and did it increase during the COVID-19 pandemic? A systematic review,” J. Academic Ethics, vol. 22, no. 2, pp. 323-343, 2024.
T. T. T. Phung, N. K. Duong, and P. L. C. Nguyen, “Recognizing cheating behavior in exams from camera images based on YOLOv8 and ResNet-CBAM models,” TNU J. Sci. Technol., vol. 230, no. 15, pp. 178-185, 2025.
T. Jannat, H. Ullah, M. Hosen, N. A. Omar, S. Arefin, and J. Hazelton, “Cracking the code: Unveiling e-cheating trends among tertiary education students,” Tertiary Educ. Manag., vol. 31, no. 4, pp. 1-25, 2026.
L. Wang et al., “Temporal segment networks: Towards good practices for deep action recognition,” in Proc. ECCV, 2016, pp. 20-36.
D. Q. Vu and T. T. T. Phung, “Simultaneous context and motion learning in video prediction,” Signal Image Video Process., vol. 17, no. 8, pp. 3933-3942, 2023.
D. Tran, L. Bourdev, R. Fergus, L. Torresani, and M. Paluri, “Learning spatiotemporal features with 3D convolutional networks,” in Proc. ICCV, 2015, pp. 4489-4497.
J. Carreira and A. Zisserman, “Quo vadis, action recognition? A new model and the kinetics dataset,” in Proc. CVPR, 2017, pp. 6299-6308.
T. Phung, V. T. Nguyen, T. H. T. Ma, and D. Q. Vu, “A (2+ 1) d attention convolutional neural network for video prediction,” in International Conference on Artificial Intelligence and Big Data in Digital Era, 2021, pp. 395-406.
C. Feichtenhofer, H. Fan, J. Malik, and K. He, “SlowFast networks for video recognition,” in Proc. ICCV, 2019, pp. 6202-6211.
A. G. Howard et al., “MobileNets: Efficient convolutional neural networks for mobile vision applications,” arXiv preprint arXiv:1704.04861, 2017, doi: 10.48550/arXiv.1704.04861.
Y. Wang, L. Chen, and R. Zhang, “TSM-MobileNetV3: A novel lightweight network model for video action recognition,” IEEE Access, vol. 11, pp. 88123-88135, 2023.
Z. Tan, W. Li, and J. Chen, “MNv3-MFAE: A lightweight network for video action recognition,” Electronics, vol. 14, no. 5, 2025, Art. no. 981.
J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proc. CVPR, 2018, pp. 7132–7141.
S. Woo, J. Park, J.-Y. Lee, and I. S. Kweon, “CBAM: Convolutional block attention module,” in Proc. ECCV, 2018, pp. 3-19.
D. Q. Vu, N. T. H. Le, and J. C. Wang, “(2+ 1) d distilled shufflenet: A lightweight unsupervised distillation network for human action recognition,” In 26th International Conference on Pattern Recognition (ICPR), 2022, pp. 3197-3203.
D. Q. Vu, T. T. H. Nguyen, D. V. Nguyen, Y. Q. Nguyen, T. N. Phung, and T. T. T. Phung, “A Lightweight Attention Model for Face Recognition,” In International Conference on Advances in Information and Communication Technology, 2023, pp. 219-227.
