Biologically Inspired Neural Networks for Crowd Escape Detection in Complex Scenes

doi:10.19678/j.issn.1000-3428.0252073

Abstract

Abstract:

Crowd escape behavior in public places is highly likely to cause serious public safety disasters. Traditional computer vision technology can detect a few characteristics of crowd escape behavior, but it is difficult to face complex dynamic visual scenes. To address this issue, based on the structural characteristics of the locust visual nerve, and leveraging the danger perception mechanism of the locust Lobula Giant Movement Detector (LGMD) as well as the mammalian retinal luminance adaptation mechanism, this paper proposes an Enhanced Crowd Escape Detection Neural Network (ECEDNN). First, the proposed neural network collects the luminance changes caused by crowd activities in the field of view. With the help of the mammalian retinal luminance adaptive mechanism, the visual response excitation is tuned to adapt to the lighting scene. Visual excitation and suppression are combined to filter background noise, and a center-surround mechanism is used to enhance motion edges. Finally, neural spike adaptive tuning is used to detect the burst escape behavior of the crowd and output strong membrane potential excitation. The experimental results show that ECEDNN can effectively detect and warn of crowd escape behavior in complex scenes, with an average accuracy of 98.90% on multiple video datasets. This work is involved the research of crowd activity detection inspired by biological visual perception mechanism, which can provide new ideas and methods for crowd behavior activity perception and anomaly detection in artificial intelligence.

Key words: anomaly detection, luminance adaptive mechanism, locust Lobula Giant Movement Detector (LGMD), visual perception, escape behavior, crowd activity, intelligent surveillance

摘要：

公共场所人群逃逸行为极易引发严重的公共安全灾难事故, 传统计算机视觉技术能检测其少许特征, 但面对复杂动态视觉场景则难以应对。针对该问题, 基于蝗虫视觉神经结构特性, 借助蝗虫小叶巨型运动检测器(LGMD)危险感知机理、哺乳动物视网膜流明自适应机制, 提出一种增强型人群逃逸检测神经网络(ECEDNN)。所提出的神经网络首先采集视野域中人群活动引发的流明变化; 然后借助哺乳动物视网膜流明自适应机制, 调谐视觉响应兴奋以适应光照场景; 接着视觉兴奋与抑制混合过滤背景噪声并采用中心环绕机制增强运动边缘; 最后神经尖峰自适应调谐用于实现对人群突发逃逸行为的检测并对其输出强烈膜电位兴奋。实验结果表明, ECEDNN能有效检测并预警复杂场景中的人群逃逸行为, 在多个视频数据集上的平均准确率达到98.90%。本文涉及生物视感机制启发的人群活动检测研究, 可为人工智能中的人群行为活动感知、异常检测等提供新思路、新方法。

关键词: 异常检测, 流明自适应机制, 蝗虫小叶巨型运动检测器, 视觉感知, 逃逸行为, 人群活动, 智能监控

FENG Tao, HU Bin, XU Guangyuan. Biologically Inspired Neural Networks for Crowd Escape Detection in Complex Scenes[J]. Computer Engineering, 2026, 52(9): 176-192.

冯涛, 胡滨, 徐光源. 复杂场景下的生物启发人群逃逸检测神经网络[J]. 计算机工程, 2026, 52(9): 176-192.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0252073

https://www.ecice06.com/EN/Y2026/V52/I9/176

Figures/Tables 18

Fig.1 Structure of ECEDNN

Fig.2 Luminance adaptive tuning process

Fig.3 Sampled frames for validity test

Fig.4 Visualization of membrane potentials in neural layer Lo

Fig.5 Output curves 1 of neuron CEBD

Fig.6 Sampled frames for response preference test

Fig.7 Output curves 2 of neuron CEBD

Fig.8 Sampled frames for ablation experiment

Fig.9 Comparison of output curves between CEBDNN and ECEDNN

Fig.10 LGMD homology models output curve

Fig.11 Sampled frames for scene interference test

References 52

1	LIANG H Y , LEE S , SUN J , et al. Unraveling the causes of the Seoul Halloween crowd-crush disaster. PLoS One, 2024, 19 (7): e0306764. doi: 10.1371/journal.pone.0306764
2	许敏, 胡滨. 密集场景下的人群拥挤检测研究综述. 计算机工程, 2026, 52 (3): 79- 96. doi: 10.19678/j.issn.1000-3428.0069340
	XU M , HU B . Survey of research on crowd congestion detection in dense scenarios. Computer Engineering, 2026, 52 (3): 79- 96. doi: 10.19678/j.issn.1000-3428.0069340
3	TYAGI B , NIGAM S , SINGH R . A review of deep learning techniques for crowd behavior analysis. Archives of Computational Methods in Engineering, 2022, 29 (7): 5427- 5455. doi: 10.1007/s11831-022-09772-1
4	PANG G S , SHEN C H , CAO L B , et al. Deep learning for anomaly detection: a review. ACM Computing Surveys, 2022, 54 (2): 1- 38.
5	SAMAILA Y A , SEBASTIAN P , SINGH N S S , et al. Video anomaly detection: a systematic review of issues and prospects. Neurocomputing, 2024, 591, 127726. doi: 10.1016/j.neucom.2024.127726
6	MEHRAN R, OYAMA A, SHAH M. Abnormal crowd behavior detection using social force model[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2009: 935-942.
7	WANG L J, DONG M. Real-time detection of abnormal crowd behavior using a matrix approximation-based approach[C]//Proceedings of the 19th IEEE International Conference on Image Processing. Washington D.C., USA: IEEE Press, 2013: 2701-2704.
8	SHEHAB D , AMMAR H . Statistical detection of a panic behavior in crowded scenes. Machine Vision and Applications, 2019, 30 (5): 919- 931. doi: 10.1007/s00138-018-0974-3
9	DIREKOGLU C, SAH M, O'CONNOR N E. Abnormal crowd behavior detection using novel optical flow-based features[C]//Proceedings of the 14th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). Washington D.C., USA: IEEE Press, 2017: 1-6.
10	SINGH G , KHOSLA A , KAPOOR R . Crowd escape event detection via pooling features of optical flow for intelligent video surveillance systems. International Journal of Image, Graphics and Signal Processing, 2019, 11 (10): 40- 49. doi: 10.5815/ijigsp.2019.10.06
11	黄昆, 齐肇建, 王娟敏, 等. 基于改进YOLOv8的密集行人检测模型. 计算机工程, 2025, 51 (5): 133- 142. doi: 10.19678/j.issn.1000-3428.0069026
	HUANG K , QI Z J , WANG J M , et al. Aggregation pedestrian detection model based on improved YOLOv8. Computer Engineering, 2025, 51 (5): 133- 142. doi: 10.19678/j.issn.1000-3428.0069026
12	YUE S G, RIND F C. A collision detection system for a mobile robot inspired by the locust visual system[C]//Proceedings of the 2005 IEEE International Conference on Robotics and Automation. Washington D.C., USA: IEEE Press, 2005: 3832-3837.
13	YUE S G , RIND F C . Collision detection in complex dynamic scenes using an LGMD-based visual neural network with feature enhancement. IEEE Transactions on Neural Networks, 2006, 17 (3): 705- 716. doi: 10.1109/TNN.2006.873286
14	HU B , YUE S G , ZHANG Z H . A rotational motion perception neural network based on asymmetric spatiotemporal visual information processing. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28 (11): 2803- 2821. doi: 10.1109/TNNLS.2016.2592969
15	刘倡, 胡滨. 生物启发的人群突发局部聚集感知神经网络. 计算机工程与应用, 2022, 58 (16): 164- 174.
	LIU C , HU B . Bio-inspired neural network for perceiving suddenly localized crowd gathering. Computer Engineering and Applications, 2022, 58 (16): 164- 174.
16	ZHAO Z X, HU B. LGMD-based neural network for detecting abnormal velocity targets in moving crowd[C]//Proceedings of the 9th International Symposium on Computer and Information Processing Technology (ISCIPT). Washington D.C., USA: IEEE Press, 2024: 562-566.
17	HU B, ZHANG Z H, LI L. LGMD-based visual neural network for detecting crowd escape behavior[C]//Proceedings of the 5th IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS). Washington D.C., USA: IEEE Press, 2019: 772-778.
18	LAMB T D . Why rods and cones?. Eye, 2016, 30 (2): 179- 185. doi: 10.1038/eye.2015.236
19	GRIMES W N , SONGCO-AGUAS A , RIEKE F . Parallel processing of rod and cone signals: retinal function and human perception. Annual Review of Vision Science, 2018, 4, 123- 141. doi: 10.1146/annurev-vision-091517-034055
20	WU S , WONG H S , YU Z W . A Bayesian model for crowd escape behavior detection. IEEE Transactions on Circuits and Systems for Video Technology, 2013, 24 (1): 85- 98.
21	WANG M D , CHANG F L , ZHANG Y M . Crowd escape event detection based on direction-collectiveness model. KSII Transactions on Internet and Information Systems (TIIS), 2018, 12 (9): 4355- 4374.
22	ALDISSI B , AMMAR H . Real-time frequency-based detection of a panic behavior in human crowds. Multimedia Tools and Applications, 2020, 79 (33): 24851- 24871.
23	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): 57. doi: 10.1007/s00138-021-01182-w
24	FAROOQ M U , SAAD M N M , KHAN S D . Motion-shape-based deep learning approach for divergence behavior detection in high-density crowd. The Visual Computer, 2022, 38 (5): 1553- 1577. doi: 10.1007/s00371-021-02088-4
25	JOSHI K V , PATEL N M . Anomaly detection in surveillance scenes using autoencoders. SN Computer Science, 2023, 4 (6): 804. doi: 10.1007/s42979-023-02260-8
26	SHARIF M H , JIAO L , OMLIN C W . Deep crowd anomaly detection: state-of-the-art, challenges, and future research directions. Artificial Intelligence Review, 2025, 58 (5): 139. doi: 10.1007/s10462-024-11092-8
27	SHARIFANI K , AMINI M . Machine learning and deep learning: a review of methods and applications. World Information Technology and Engineering Journal, 2023, 10 (7): 3897- 3904.
28	ROKA S, DIWAKAR M, KARANWAL S. A review in anomalies detection using deep learning[C]//Proceedings of the 3rd International Conference on Sustainable Computing. Singapore: Springer, 2022: 329-338.
29	IDREES S , MANOOKIN M B , RIEKE F , et al. Biophysical neural adaptation mechanisms enable artificial neural networks to capture dynamic retinal computation. Nature Communications, 2024, 15, 5957. doi: 10.1038/s41467-024-50114-5
30	GRIFFIS K G , FEHLHABER K E , RIEKE F , et al. Light adaptation of retinal rod bipolar cells. The Journal of Neuroscience, 2023, 43 (24): 4379- 4389. doi: 10.1523/JNEUROSCI.0444-23.2023
31	GLORIANI A H , SCHÜTZ A C . Humans trust central vision more than peripheral vision even in the dark. Current Biology, 2019, 29 (7): 1206- 1210. doi: 10.1016/j.cub.2019.02.023
32	LOPEZ-HAZAS J , MONTERO A , RODRIGUEZ F B . Influence of bio-inspired activity regulation through neural thresholds learning in the performance of neural networks. Neurocomputing, 2021, 462, 294- 308. doi: 10.1016/j.neucom.2021.08.001
33	FONTAINE B , PEÑA J L , BRETTE R . Spike-threshold adaptation predicted by membrane potential dynamics in vivo. PLoS Computational Biology, 2014, 10 (4): e1003560. doi: 10.1371/journal.pcbi.1003560
34	HUANG C , RESNIK A , CELIKEL T , et al. Adaptive spike threshold enables robust and temporally precise neuronal encoding. PLoS Computational Biology, 2016, 12 (6): e1004984. doi: 10.1371/journal.pcbi.1004984
35	HOMBERG U , BRANDL C , CLYNEN E , et al. Mas-allatotropin/Lom-AG-myotropin I immunostaining in the brain of the locust, Schistocerca gregaria. Cell and Tissue Research, 2004, 318 (2): 439- 457. doi: 10.1007/s00441-004-0913-7
36	ZHU Y , DEWELL R B , WANG H X , et al. Pre-synaptic muscarinic excitation enhances the discrimination of looming stimuli in a collision-detection neuron. Cell Reports, 2018, 23 (8): 2365- 2378. doi: 10.1016/j.celrep.2018.04.079
37	WERNITZNIG S , RIND F C , ZANKEL A , et al. The complex synaptic pathways onto a looming-detector neuron revealed using serial block-face scanning electron microscopy. Journal of Comparative Neurology, 2022, 530 (2): 518- 536. doi: 10.1002/cne.25227
38	HU B , ZHANG Z H . Bio-inspired visual neural network on spatio-temporal depth rotation perception. Neural Computing and Applications, 2021, 33 (16): 10351- 10370. doi: 10.1007/s00521-021-05796-z
39	KAWAI F . Certain retinal horizontal cells have a center-surround antagonistic organization. Journal of Neurophysiology, 2022, 128 (5): 1337- 1343. doi: 10.1152/jn.00163.2022
40	CABELLO-SOLORZANO K, DE ARAUJO I O, PEÑA M, et al. The impact of data normalization on the accuracy of machine learning algorithms: a comparative analysis[C]//Proceedings of the 18th International Conference on Soft Computing Models in Industrial and Environmental Applications. Berlin, Germany: Springer, 2023: 344-353.
41	FERRYMAN J, SHAHROKNI A. PETS2009: dataset and challenge[C]//Proceedings of the 12th IEEE International Workshop on Performance Evaluation of Tracking and Surveillance. Washington D.C., USA: IEEE Press, 2009: 1-6.
42	RABIEE H, HADDADNIA J, MOUSAVI H, et al. Novel dataset for fine-grained abnormal behavior understanding in crowd[C]//Proceedings of the 13th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). Washington D.C., USA: IEEE Press, 2016: 95-101.
43	DEGARDIN B, PROENÇA H. Human activity analysis: iterative weak/self-supervised learning frameworks for detecting abnormal events[C]//Proceedings of the IEEE International Joint Conference on Biometrics (IJCB). Washington D.C., USA: IEEE Press, 2020: 1-7.
44	CHAN A B , VASCONCELOS N . Modeling, clustering, and segmenting video with mixtures of dynamic textures. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, 30 (5): 909- 926. doi: 10.1109/TPAMI.2007.70738
45	YUE S G , RIND F C . Redundant neural vision systems—competing for collision recognition roles. IEEE Transactions on Autonomous Mental Development, 2013, 5 (2): 173- 186. doi: 10.1109/TAMD.2013.2255050
46	张娓娓, 陈绥阳, 陈锐. 视频监控下利用改进型C3D-RF的人群异常行为检测. 光学技术, 2021, 47 (2): 187- 195.
	ZHANG W W , CHEN S Y , CHEN R . Abnormal crowd behavior detection using improved C3D-RF under Video Surveillance. Optical Technique, 2021, 47 (2): 187- 195.
47	DENGXIONG X, BAO W T, KONG Y. Multiple instance relational learning for video anomaly detection[C]//Proceedings of the International Joint Conference on Neural Networks (IJCNN). Washington D.C., USA: IEEE Press, 2021: 1-8.
48	TRIPATHY S K , SUDHAMSH R , SRIVASTAVA S , et al. MuST-POS: multiscale spatial-temporal 3D atrous-net and PCA guided OC-SVM for crowd panic detection. Journal of Intelligent & Fuzzy Systems, 2022, 42 (4): 3501- 3516.
49	徐桂菲, 王平, 罗凡波, 等. 基于卷积神经网络的人群突散异常行为检测. 计算机工程与设计, 2022, 43 (5): 1389- 1396.
	XU G F , WANG P , LUO F B , et al. Detection of abrupt dispersal of abnormal human behavior based on convolutional neural network. Computer Engineering and Design, 2022, 43 (5): 1389- 1396.
50	FAN Z Y , YI S H , WU D , et al. Video anomaly detection using CycleGan based on skeleton features. Journal of Visual Communication and Image Representation, 2022, 85, 103508. doi: 10.1016/j.jvcir.2022.103508
51	ALAFIF T , ALZAHRANI B , CAO Y , et al. Generative adversarial network based abnormal behavior detection in massive crowd videos: a Hajj case study. Journal of Ambient Intelligence and Humanized Computing, 2022, 13 (8): 4077- 4088. doi: 10.1007/s12652-021-03323-5
52	邢天祎, 郭茂祖, 陈加栋, 等. 基于空时对抗变分自编码器的人群异常行为检测. 智能系统学报, 2023 (5): 994- 1004.
	XING T Y , GUO M Z , CHEN J D , et al. Detection of abnormal crowd behavior based on spatial-temporal adversarial variational autoencoder. CAAI Transactions on Intelligent Systems, 2023 (5): 994- 1004.

[1]	CHEN Xin, SUN Yicheng, TAN Cheng. Survey on Log Anomaly Detection Methods Based on Large Language Models [J]. Computer Engineering, 2026, 52(8): 1-15.
[2]	ZHANG Chunhao, XIE Bin, ZHANG Jiahao. Hybrid Method of Improved VAE and BiLSTM for Unsupervised Time Series Anomaly Detection [J]. Computer Engineering, 2026, 52(7): 131-142.
[3]	ZHANG Penghe, YANG Yining, WANG Bicheng, YI Yunqi, TANG Zhongrui, LIU Min. Anomaly Detection Method for Electricity Meter Based on D-DADA Algorithm and DBE-YOLO Network [J]. Computer Engineering, 2026, 52(5): 445-455.
[4]	LIU Ze, SONG Tinglun, SHI Xianrang, SU Yang, ZHAO Qun. Research on Pavement Anomaly Detection Technology Based on Improved RT-DETR [J]. Computer Engineering, 2026, 52(4): 187-199.
[5]	DONG Fanghe, SHI Qiong, SHI Zhibin. Adversarial Traffic Detection Method Based on Ensemble Learning and Anomaly Detection [J]. Computer Engineering, 2026, 52(2): 275-286.
[6]	XU Shipeng, WANG Lei, SHENG Jie. Knowledge Graph-based Advance Recognition Model for Abnormal Individuals [J]. Computer Engineering, 2025, 51(9): 59-70.
[7]	WU Jiehui, LIU Yi. Jointly Optimized Anomaly Detection Model for Multivariate Time Series [J]. Computer Engineering, 2025, 51(9): 166-176.
[8]	WANG Guangming, LI Dongqing, JIANG Congfeng. Network Traffic Anomaly Detection for Data Centers in Imbalanced Datasets [J]. Computer Engineering, 2025, 51(8): 227-237.
[9]	ZHANG Anqin, DING Zhifeng. Network Anomaly Detection Integrating Dynamic Graph Embedding and Transformer Autoencoder [J]. Computer Engineering, 2025, 51(4): 47-56.
[10]	NIU Xuanwen, DU Ye, YANG Mingsong, LI Ang, LI Meihong. Anomaly Detection Method for Satellite Networks Based on Hierarchical Federation and Dynamic Weights [J]. Computer Engineering, 2025, 51(12): 210-220.
[11]	DI Qinbo, CHEN Shaoli, SHI Liangren. GRD: Anomaly Detection Algorithm for Multivariate Time Series Data Based on GNN and Diffusion Model [J]. Computer Engineering, 2025, 51(11): 35-44.
[12]	Xuanyu ZHOU, Lianhua WU, Qinhua ZHENG, Tianxing XIAO, Zixuan WANG, Simin ZHANG. Weakly Supervised Video Anomaly Detection Method for Rope Skipping Combined with Semantic Prompts and Memory Enhancement [J]. Computer Engineering, 2024, 50(7): 87-95.
[13]	LI Yongfei, LI Mingyang, CHANG Xin, CAO Kexin. Anomaly Detection of IoT Water Quality Monitoring Data Based on Explainable Deep Learning [J]. Computer Engineering, 2024, 50(6): 179-187.
[14]	Bochao ZHAO, Jiajun MA, Lei CUI, Wenpeng LUAN, Jing ZHU. Anomaly Detection for Photovoltaic Based on Improved VMD-XGBoost-BiLSTM Combination Model [J]. Computer Engineering, 2024, 50(3): 306-316.
[15]	Jun WANG, Huixia LAI, Yue WAN, Shi ZHANG. Angle-based Graph Neural Network Method for Anomaly Detection in High Dimensional Data [J]. Computer Engineering, 2024, 50(3): 156-165.

Please choose a citation manager

Content to export