STFE-Net:Enhanced Deep Learning for Real-Time Abnormal Behavior Detection in Video Surveillance
Abstract
A spatiotemporal feature enhancement network (STFE-Net) is proposed for real-time abnormal behavior detection in surveillance video. The model integrates optimized 3D convolution kernels, a multi-scale convolution strategy, and a feature fusion module with skip connections and attention mechanisms, followed by an improved fully connected classification structure. STFE-Net achieves an average accuracy of 0.85 on the UCF-Crime dataset and 0.82 on multiple datasets, outperforming traditional 3D-CNNs (average 0.65) and RNN-based models (average 0.68). False alarm rates are reduced to 0.10 and false negative rates to 0.07, demonstrating improved precision and robustness. Compared with baseline methods, STFE-Net shows a 26.2% increase in average accuracy and up to 50% reduction in false positives, significantly improving real-time surveillance reliability.References
.Umale-Nagmote A, Goel C, Lal N. Enhanced intelligent video monitoring using hybrid integration of spatiotemporal autoencoders and convolutional LSTMs. Online-only issue. 2025;49(18).
.Aberkane S, Elarbi-Boudihir M. Deep reinforcement learning-based anomaly detection for video surveillance. Online-only issue. 2022;46(2).
.Alarcon JA, Santamaria F, Al-Sumaiti AS, Rivera S. Low-Capacity Exploitation of Distribution Networks and Its Effect on the Planning of Distribution Networks. Energies. 2020;13(8). DOI: 10.3390/en13081920
.Zhang YX, Song JC, Jiang YH, Li HJ. Online Video Anomaly Detection. Sensors. 2023;23(17). DOI: 10.3390/s23177442
.Chen W, Yu ZH, Yang CC, Lu YY. Abnormal Behavior Recognition Based on 3D Dense Connections. International Journal of Neural Systems. 2024;34(09). DOI: 10.1142/s0129065724500497
.Li JN, Zhou BR, Yao WF, Zhao WM, Cheng RL, Ou MY, et al. Research on dynamic robust planning method for active distribution network considering correlation. Frontiers in Energy Research. 2023;11. DOI: 10.3389/fenrg.2023.1338136
.Chen D, Zhang S. Deep learning-based involution feature extraction for human posture recognition in martial arts. Online-only issue. 2025;49(12).
.Li JN, Wang T, Tang SQ, Jiang JR, Chen SY. Planning distribution network using the multi-agent game and distribution system operators. Frontiers in Energy Research. 2023;11. DOI: 10.3389/fenrg.2023.1244394
.Le T, Huynh-Duc N, Nguyen CT, Tran MT. Motion embedded images: An approach to capture spatial and temporal features for action recognition. Online-only issue. 2023;47(3).
.Chang CW, Chang CY, Lin YY. A hybrid CNN and LSTM-based deep learning model for abnormal behavior detection. Multimedia Tools and Applications. 2022;81(9):11825-43. DOI: 10.1007/s11042-021-11887-9
.Wu CF, Cheng ZX. A Novel Detection Framework for Detecting Abnormal Human Behavior. Mathematical Problems in Engineering. 2020;2020. DOI: 10.1155/2020/6625695
Sugianto N, Tjondronegoro D, Sorwar G. Collaborative federated learning framework to minimize data transmission for AI-enabled video surveillance. Information Technology & People. 2025;38(3):1526–50. doi:10.1108/ITP-08-2021-0598.
Lysova T. Intersecting perspectives: Video surveillance in urban spaces through surveillance society and security state frameworks. Cities. 2025;156:105544. doi:10.1016/j.cities.2024.105544.
.Li ZH, Wu WC, Zhang BM, Tai X. Feeder-corridor-based distribution network planning model with explicit reliability constraints. Iet Generation Transmission & Distribution. 2020;14(22):5310-8. DOI: 10.1049/iet-gtd.2020.1093
.Duja KU, Khan IA, Alsuhaibani M. Video Surveillance Anomaly Detection: A Review on Deep Learning Benchmarks. Ieee Access. 2024;12:164811-42. DOI: 10.1109/access.2024.3491868
.Deng LJ, Fu RC, Sun Q, Jiang M, Li ZH, Chen H, et al. Abnormal behavior recognition based on feature fusion C3D network. Journal of Electronic Imaging. 2023;32(2). DOI: 10.1117/1.Jei.32.2.021605
.Ardabili BR, Pazho AD, Noghre GA, Katariya V, Hull G, Reid S, Tabkhi H. Exploring public's perception of safety and video surveillance technology: A survey approach. Technology in Society. 2024;78:102641. doi:10.1016/j.techsoc.2024.102641. DOI: 10.17775/cseejpes.2021.08540
.Lv ZH, Xiao J, He GW, Jiao H, Zhou YP. A planning method based on total supply capability for active distribution network. Energy Reports. 2025;13:973-86. DOI: 10.1016/j.egyr.2024.12.056
.Lovecek T, Boros M, Mäkká K, Maris L. Designing of intelligent video-surveillance systems in road tunnels using software tools. Sustainability. 2023;15(7):5702. doi:10.3390/su15075702.
.Ammar H, Cherif A. DeepROD: a deep learning approach for real-time and online detection of a panic behavior in human crowds. Machine Vision and Applications. 2021;32(3). DOI: 10.1007/s00138-021-01182-w
.Teall AM, Bobek H, Zeno R, Graham MC. An innovative well-child video project to teach developmental surveillance and anticipatory guidance. Journal of Nursing Education. 2023;62(7):412–5. doi:10.3928/01484834-20230315-03.
.Liang PH, Li XJ, Guo YC. Local governments and the diffusion of video surveillance in China: Evidence from the public procurement contracts. Journal of Chinese Political Science. 2025. doi:10.1007/s11366-025-09916-7.
DOI:
https://doi.org/10.31449/inf.v49i23.10130Downloads
Published
How to Cite
Issue
Section
License
I assign to Informatica, An International Journal of Computing and Informatics ("Journal") the copyright in the manuscript identified above and any additional material (figures, tables, illustrations, software or other information intended for publication) submitted as part of or as a supplement to the manuscript ("Paper") in all forms and media throughout the world, in all languages, for the full term of copyright, effective when and if the article is accepted for publication. This transfer includes the right to reproduce and/or to distribute the Paper to other journals or digital libraries in electronic and online forms and systems.
I understand that I retain the rights to use the pre-prints, off-prints, accepted manuscript and published journal Paper for personal use, scholarly purposes and internal institutional use.
In certain cases, I can ask for retaining the publishing rights of the Paper. The Journal can permit or deny the request for publishing rights, to which I fully agree.
I declare that the submitted Paper is original, has been written by the stated authors and has not been published elsewhere nor is currently being considered for publication by any other journal and will not be submitted for such review while under review by this Journal. The Paper contains no material that violates proprietary rights of any other person or entity. I have obtained written permission from copyright owners for any excerpts from copyrighted works that are included and have credited the sources in my article. I have informed the co-author(s) of the terms of this publishing agreement.
Copyright © Slovenian Society Informatika
