The DOI matches its Crossref record; 23 of the 25 references it cites were found in a registry or an index.
You'll never walk alone: Modeling social behavior for multi-target tracking
Abstract
Object tracking typically relies on a dynamic model to predict the object's location from its past trajectory. In crowded scenarios a strong dynamic model is particularly important, because more accurate predictions allow for smaller search regions, which greatly simplifies data association. Traditional dynamic models predict the location for each target solely based on its own history, without taking into account the remaining scene objects. Collisions are resolved only when they happen. Such an approach ignores important aspects of human behavior: people are driven by their future destination, take into account their environment, anticipate collisions, and adjust their trajectories at an early stage in order to avoid them. In this work, we introduce a model of dynamic social behavior, inspired by models developed for crowd simulation. The model is trained with videos recorded from birds-eye view at busy locations, and applied as a motion model for multi-people tracking from a vehicle-mounted camera. Experiments on real sequences show that accounting for social interactions and scene knowledge improves tracking performance, especially during occlusions.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Histograms of Oriented Gradients for Human Detection
Dalal, N., & Triggs, B. (2005) · 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) · doi:10.1109/cvpr.2005.177 · Cited by 33,584
Social force model for pedestrian dynamics.
Helbing, D., & Molnár, P. (1995) · Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics · doi:10.1103/physreve.51.4282 · Cited by 7,299
A discriminatively trained, multiscale, deformable part model
Felzenszwalb, P. F., McAllester, D., & Ramanan, D. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587597 · Cited by 2,972
Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics
Bernardin, K., & Stiefelhagen, R. (2008) · EURASIP Journal on Image and Video Processing · doi:10.1155/2008/246309 · Cited by 2,918
Abnormal crowd behavior detection using social force model
Mehran, R., Oyama, A., & Shah, M. A. (2009) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2009.5206641 · Cited by 1,698
Lerner, A., Chrysanthou, Y., & Lischinski, D. (2007) · Computer Graphics Forum · doi:10.1111/j.1467-8659.2007.01089.x · Cited by 1,199
A Boosted Particle Filter: Multitarget Detection and Tracking
Okuma, K., Taleghani, A., de Freitas, N., Little, J. J., & Lowe, D. G. (2004) · Lecture notes in computer science · doi:10.1007/978-3-540-24670-1_3 · Cited by 1,025
Gräbner, H., & Bischof, H. (2006) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2006.215 · Cited by 998
Global data association for multi-object tracking using network flows
Zhang, L., Li, Y., & Nevatia, R. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587584 · Cited by 945
People-tracking-by-detection and people-detection-by-tracking
Andriluka, M., Roth, S., & Schiele, B. (2008) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2008.4587583 · Cited by 825
Flagged first, then by citations
2 of the 25 references couldn’t be matched to a known paper.
Research works printed but not linked: 1
- schadschneider, 2001, Cellular automaton approach to pedestrian dynamics—theory, PED
Cited by
Sorted by citations
Items 601–700 of 1,000
AdapTraj: A Multi-Source Domain Generalization Framework for Multi-Agent Trajectory Prediction
Qian, T.-W., Chen, Y.-L., Cong, G., Xu, Y., & Wang, F. (2023) · IEEE International Conference on Data Engineering · doi:10.1109/icde60146.2024.00113 · Cited by 21
“The individual specific extractor M ind is designed to extract specific features H si from the individual’s mobility, while the neighbor specific extractor M nei focuses on extracting specific features H s E i from the neighbors’ interaction.”
BackgroundWang, R., Lam, S.-K., Wu, M., et al. (2024) · Measurement · doi:10.1016/j.measurement.2024.115470 · Cited by 18
Recurrent Aligned Network for Generalized Pedestrian Trajectory Prediction
Dong, Y., Wang, L., Zhou, S., Hua, G., & Sun, C. (2024) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2024.3493966 · Cited by 18
“The ETH-UCY dataset [64], [65] is captured in the scenes of outdoor pedestrian paths.”
Pedestrian action recognition using motion type classification
Hariyono, J., & Jo, K.-H. (2015) · IEEE International Conference on Cybernetics (CYBCONF) · doi:10.1109/cybconf.2015.7175919 · Cited by 18
“Probabilistic path prediction is used to classify whether pedestrian will cross or stopping [9], [11].”
MethodClosed-Loop Adaptation for Robust Tracking
Fan, J., Shen, X., & Wu, Y. N. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15549-9_30 · Cited by 18
“At the higher level, the interactions between different targets are explored in multiple target tracking, either by a Markov network [5] or by modeling their social behaviors [4].”
BackgroundData Association for Multi-Object Visual Tracking
Betke, M., & Wu, Z. (2017) · Synthesis lectures on computer vision · doi:10.1007/978-3-031-01816-9 · Cited by 18
Group Split and Merge Prediction With 3D Convolutional Networks
Wang, A., & Steinfeld, A. M. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2969947 · Cited by 18
“The ETH dataset [26] contains two sets of data from two scenes (ETH and HOTEL).”
BackgroundMethodPedestrian behavior prediction model with a convolutional LSTM encoder–decoder
Chen, K., Song, X., Han, D., et al. (2020) · Physica A Statistical Mechanics and its Applications · doi:10.1016/j.physa.2020.125132 · Cited by 18
Niu, Z., Jia, X., & Yao, W. (2022) · IEEE Access · doi:10.1109/access.2022.3148145 · Cited by 18
“In addition, there are methods such as Bayesian formulation [17], interactive Gaussian Processes (IGP) [18] and Linear Trajectory Avoidance (LTA) [19], etc.”
MethodA motion model based on recurrent neural networks for visual object tracking
Shahbazi, M., Bayat, M. H., & Tarvirdizadeh, B. (2022) · Image and Vision Computing · doi:10.1016/j.imavis.2022.104533 · Cited by 19
Samavi, S., Han, J. R., Shkurti, F., et al. (2024) · IEEE Transactions on Robotics · doi:10.1109/tro.2024.3484634 · Cited by 18
“Second, to justify the use of ORCA as a model of pedestrians, we validate its trajectory forecasting performance on real human data in the commonly-used ETH/UCY benchmark [25], [26].”
MethodDiffWT: Diffusion-Based Pedestrian Trajectory Prediction With Time-Frequency Wavelet Transform
Chen, X., Zeng, L., Gao, M., Ding, C., & Bian, Y. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3489638 · Cited by 20
“3) Extensive experiments are conducted on two benchmark trajectory prediction datasets: 1) the ETH/UCY dataset [26], [27] and 2) the Stanford drone dataset (SDD) [28].”
Unified Spatial–Temporal Edge-Enhanced Graph Networks for Pedestrian Trajectory Prediction
Li, R., Qiao, T., Katsigiannis, S., Zhu, Z., & Shum, H. P. H. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3539522 · Cited by 22
“Our approach outperforms state-of-the-art methods on commonly used pedestrian trajectory prediction datasets, including ETH [30], UCY [31] and Stanford Drone Dataset (SDD) [32].”
Learning the Pedestrian-Vehicle Interaction for Pedestrian Trajectory Prediction
Zhang, C., & Berger, C. (2022) · International Conference on Control, Automation and Robotics (ICCAR) · doi:10.1109/iccar55106.2022.9782673 · Cited by 17
“Besides, the ETH [13] and UCY [14] datasets, which are commonly used to evaluate pedestrian trajectory predictions, do not include annotations for vehicles, and hence, are not suitable for an appropriate evaluation of this task.”
BackgroundRuan, K., & Di, X. S. (2024) · Computers · doi:10.3390/computers13060151 · Cited by 17
Chen, X., Liang, Y., Hu, C., et al. (2025) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2025.3540929 · Cited by 17
“To validate the effectiveness of the proposed method, we utilize four publicly available trajectory datasets, namely ETH [52], UCY [53], Stanford Drones Dataset (SDD) [54], and INTERACTION [55].”
Detecting and Tracking Sports Players with Random Forests and Context-Conditioned Motion Models
Liu, J., & Carr, P. (2014) · Advances in computer vision and pattern recognition · doi:10.1007/978-3-319-09396-3_6 · Cited by 17
“Recently, multi-object motion models have been used in pedestrian tracking to anticipate how people will change their trajectories to avoid collisions [20], or for estimating whether a pair of trajectories have correlated motions [4].”
BackgroundDetecting Social Groups in Crowded Surveillance Videos Using Visual Attention
Leach, M., Baxter, R. H., Robertson, N. M., et al. (2014) · IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2014.75 · Cited by 17
“improve tracking [1] and behavior analysis.”
Profiling stationary crowd groups
Yi, S., & Wang, X. (2014) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2014.6890138 · Cited by 17
“Recently the social force model [12] and other agentbased models have been applied to behavior analysis in computer vision [13, 14].”
MethodRobust individual and holistic features for crowd scene classification
Liu, W., Lau, R. W. H., & Manocha, D. (2016) · Pattern Recognition · doi:10.1016/j.patcog.2016.03.031 · Cited by 17
Beyond Group: Multiple Person Tracking via Minimal Topology-Energy-Variation
Gao, S., Ye, Q., Xing, J., Kuijper, A., Han, Z., Jiao, J., & Ji, X. (2017) · IEEE Transactions on Image Processing · doi:10.1109/tip.2017.2708901 · Cited by 17
“How to reach a strong local minimum being not far away from the global minimum in limited iterations has been investigated in many previous work [7], [8], [46].”
BackgroundReal-Time Certified Probabilistic Pedestrian Forecasting
Jacobs, H. O., Hughes, O., Johnson‐Roberson, M., et al. (2017) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2017.2719762 · Cited by 17
“ance through an interaction potential [11]–[13].”
A comprehensive survey on trajectory-based location prediction
Koolwal, V., & Mohbey, K. K. (2020) · Iran Journal of Computer Science · doi:10.1007/s42044-019-00052-z · Cited by 17
“The proposed novel model predicts trajectories of a pedestrian more accurately (outperforms) than state-of-the-art methods on two standard datasets, ETH [55] and UCY [56].”
MethodDimensionally Aware Multi-Objective Genetic Programming for Automatic Crowd Behavior Modeling
Li, D., & Zhong, J. (2020) · ACM Transactions on Modeling and Computer Simulation · doi:10.1145/3391407 · Cited by 17
“Pellgrini et al. [25] proposed a model called linear trajectory avoidance (LTA) for pedestrians with short-term prediction.”
BackgroundAutoTrajectory: Label-Free Trajectory Extraction and Prediction from Videos Using Dynamic Points
Ma, Y., Zhu, X., Cheng, X., Yang, R., Liu, J., & Manocha, D. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58601-0_38 · Cited by 17
“One general method to get a trajectory dataset [35,26] is to label consecutive positions of moving traffic-agents (pedestrians or vehicles) on fixed-view videos and then transfer the trajectory from the image coordinate system to a real-world coordinate system.”
MethodHuman Trajectory Prediction Using Similarity-Based Multi-Model Fusion
Habibi, G., & How, J. P. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3048652 · Cited by 17
“[10]: defined as the average Euclidean distance (L2 norm) between the predicted and ground truth trajectory over a fixed time horizon T :”
BackgroundMethodREACH - Realtime crowd tracking using a hybrid motion model
Bera, A., & Manocha, D. (2015) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2015.7139261 · Cited by 17
“Other recently-developed discrete approaches are based on cognitive models [9], affordance [12], short-term planning using a discrete approach [2], linear trajectory avoidance (LTA) [29], or perceptual models [28].”
BackgroundMethodThe Social Force PHD Filter for Tracking Pedestrians
Krishnan, K., Chen, X., Tharmarasa, R., et al. (2017) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2017.2680718 · Cited by 17
Trajectories as Topics: Multi-Object Tracking by Topic Discovery
Luo, W., Stenger, B., Zhao, X., & Kim, T. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2866955 · Cited by 17
“We compare our results with those in [2], [1], [5], [7].”
BackgroundMethodResultInfluentialEgocentric Basketball Motion Planning from a Single First-Person Image
Shi, J., Chan, A., & Bertasius, G. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00617 · Cited by 25
Egocentric Basketball Motion Planning from a Single First-Person ImagePreprint
Bertasius, G., Chan, A., & Shi, J. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.01413 · Cited by 17
Yang, Q., Mei, Q., Fan, C., Ma, M., & Li, X. (2023) · Buildings · doi:10.3390/buildings13061502 · Cited by 16
Liu, Z., Na, W., Yao, C., Liu, C., & Chen, Q. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3390561 · Cited by 17
“Additionally, we use ETH Hotel dataset [22] for quantitative experiments of sim-to-real without additional training.”
MethodZhang, Y., & Wang, Y. (2025) · EURASIP Journal on Image and Video Processing · doi:10.1186/s13640-025-00677-0 · Cited by 23
A Unified Multiple-Motion-Mode Framework for Socially Compliant Navigation in Dense Crowds
Chen, Y., & Lou, Y. (2021) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2021.3125367 · Cited by 16
Unified and Fast Human Trajectory Prediction Via Conditionally Parameterized Normalizing Flow
Sun, J., Wang, Z., Li, J., et al. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3133862 · Cited by 16
Semantic scene upgrades for trajectory prediction
Syed, A., & Morris, B. (2023) · Machine Vision and Applications · doi:10.1007/s00138-022-01357-z · Cited by 16
“For ETH–UCY the metrics are in where N is the total number of trajectories.”
MethodInfluentialObstacle Avoidance Learning for Robot Motion Planning in Human–Robot Integration Environments
Hong, Y., Ding, Z., Yuan, Y., et al. (2023) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2023.3242373 · Cited by 16
“Several models that encode pedestrian intentions to predict their future trajectories have been designed manually, such as the constant speed models [8], discrete choice model [9], [10], social force model [11], and its variants [12].”
MethodZhang, Z., Hajieghrary, H., Dean‐Leon, E., et al. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3296333 · Cited by 16
“In earlier studies of dynamic obstacle avoidance, it is assumed that the motions of obstacles are known or modeled by deterministic motion models, such as the constant velocity model [16] and the reciprocal velocity obstacle [17].”
BackgroundA Systematic Analysis of Subgroup Research in Pedestrian and Evacuation Dynamics
Wu, W., & Zheng, X. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3318417 · Cited by 18
“Our survey shows that only a few studies have attempted to address this issue [118], [119], whereas the vast majority of datasets are labeled with independent pedes-trians.”
BackgroundINT2: Interactive Trajectory Prediction at Intersections
Yan, Z., Li, P., Fu, Z., Xu, S., Shi, Y., Chen, X., Zheng, Y., Li, Y., Liu, T., Li, C., et al. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00784 · Cited by 16
“Conventional datasets focus on forecasting pedestrian trajectory in campus [23, 32, 43].”
BackgroundEnhanced Crowd Dynamics Simulation with Deep Learning and Improved Social Force Model
Yan, D., Ding, G., Huang, K., et al. (2024) · Electronics · doi:10.3390/electronics13050934 · Cited by 16
“Two publicly available large-scale crowd datasets, the ETH BIWI Walking Pedestrians Dataset (ETH) [50] and the University of Cyprus Dataset (UCY) [14], are used to train the model.”
MethodSun, C., Wang, B., Leng, J., et al. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3409174 · Cited by 16
Long-Term Human Trajectory Prediction Using 3D Dynamic Scene Graphs
Gorlo, N., Schmid, L., & Carlone, L. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3482169 · Cited by 20
“Thus, trajectories in current benchmark datasets are oftentimes not longer than 10s [7], [8].”
BackgroundMethodInfluentialInteractive Crowd Content Generation and Analysis Using Trajectory-Level Behavior Learning
Kim, S., Bera, A., & Manocha, D. (2015) · IEEE International Symposium on Multimedia (ISM) · doi:10.1109/ism.2015.89 · Cited by 16
“0 in [15], our approach corresponding values are in the range 0.”
BackgroundInfluentialSocially-Aware Graph Convolutional Network for Human Trajectory Prediction
Sun, Y., He, T., Hu, J., et al. (2019) · IEEE Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · doi:10.1109/itnec.2019.8729387 · Cited by 16
“To deal with the interaction, pioneering works such as social force model [8]–[10] and energy potential model [3], [11] design parameterized dynamics to simulate human behaviors.”
BackgroundCrowd Density Forecasting by Modeling Patch-Based Dynamics
Minoura, H., Yonetani, R., Nishimura, M., & Ushiku, Y. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3043169 · Cited by 16
“across the three videos, while adding the ‘Hotel’ sequence (19 350 frames captured at 25 fps) from the ETH dataset [39]”
MethodSearch-Based Online Trajectory Planning for Car-like Robots in Highly Dynamic Environments
Lin, J., Zhou, T., Zhu, D., et al. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9560852 · Cited by 16
“We then evaluate our method on three public publicly available datasets from ETH [24] and UCY [25].”
Ratsamee, P., Mae, Y., Kamiyama, K., Horade, M., Kojima, M., & Arai, T. (2015) · ROBOMECH Journal · doi:10.1186/s40648-015-0033-z · Cited by 16
“Most research pays attention to how to integrate these factors via a number of different models [5, 8, 15].”
BackgroundPredicting Pedestrian Trajectories
Pellegrini, S., Ess, A., & Gool, L. V. (2011) · Springer eBooks · doi:10.1007/978-0-85729-997-0_24 · Cited by 16
“While in [14] we show that taking into account these social factors improves prediction and object tracking performance compared to a”
BackgroundMethodResultInfluentialSelf-citationTracking with multi-level featuresPreprint
Henschel, R. D., Leal-Taixé, L., Rosenhahn, B., & Schindler, K. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1607.07304 · Cited by 17
Self-citationGroup-based Motion Prediction for Navigation in Crowded EnvironmentsPreprint
Wang, A., Mavrogiannis, C., & Steinfeld, A. M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.11637 · Cited by 16
“We conduct an extensive empirical evaluation over 5 different human datasets [14, 15], each with a flow following and a crossing scenario.”
BackgroundMethodLi, Y., Xie, C., Liang, R., et al. (2023) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2023.3308572 · Cited by 15
“Traditional methods [6], [7], [8], [9] heavily rely on hand-crafted rules to describe human motions and human-to-human interactions.”
BackgroundMethodInfluentialA Review of Trajectory Prediction Methods for the Vulnerable Road User
Schuetz, E., & Flohr, F. B. (2023) · Robotics · doi:10.3390/robotics13010001 · Cited by 20
“Encoding the positions with a global coordinate system is typically used in static scenes, such as the surveillance scenes in the ETH/UCY datasets [106,107] or drone scenes in the Stanford Drone Dataset (SDD) [108], given the sensor’s origin never changes.”
BackgroundMethodZeng, W., Lin, Q., Zhu, B., Peng, C., & Yu, R. (2024) · Expert Systems with Applications · doi:10.1016/j.eswa.2024.123674 · Cited by 15
Sparse Transformer Network With Spatial-Temporal Graph for Pedestrian Trajectory Prediction
Gao, L., Gu, X., Chen, F., & Wang, J. (2024) · IEEE Access · doi:10.1109/access.2024.3442442 · Cited by 20
“In this section, we evaluate our methods on two publicly available datasets: ETH and UCY [40], [41], with a total of five sets of data, including four different scenarios.”
MethodTrajFine: Predicted Trajectory Refinement for Pedestrian Trajectory Forecasting
Wang, K.-L., Tsao, L.-W., Wu, J.-C., et al. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00451 · Cited by 15
“The widely recognized ETH-UCY dataset [23, 32] comprises five subsets: ETH, HOTEL, UNIV, ZARA1, and ZARA2, where the first two subsets are from ETH and the remaining belong to UCY.”
BackgroundZong, M., Chang, Y., Dang, Y., et al. (2024) · Applied Sciences · doi:10.3390/app14209349 · Cited by 15
IGGCN: Individual-guided graph convolution network for pedestrian trajectory prediction
Chen, W., Sang, H., Wang, J., et al. (2024) · Digital Signal Processing · doi:10.1016/j.dsp.2024.104862 · Cited by 15
LG-Traj: LLM Guided Pedestrian Trajectory Prediction
Chib, P. S., & Singh, P. P. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00709 · Cited by 43
“Datasets To validate our proposed approach, we conduct experiments on two benchmark datasets (Table 1): the Stanford Drone Dataset (SDD) [45] and ETH-UCY dataset [27, 44].”
Humans in groups: The importance of contextual information for understanding collective activities
Noceti, N., & Odone, F. (2014) · Pattern Recognition · doi:10.1016/j.patcog.2014.05.008 · Cited by 15
Multi-target tracking of time-varying spatial patterns
Liu, J., & Liu, Y. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5539855 · Cited by 15
“Our prior knowledge can be encoded into various models to regulate the tracking, for example, previous work has used the adaptive appearance models [19, 11], temporal motion models (filters) [9], spatiotemporal consistency models [21], target interaction models [8, 22, 16] and global scene constraint models [1, 18].”
MethodNatural vision based method for predicting pedestrian behaviour in urban environments
Vasishta, P., Vaufreydaz, D., & Spalanzani, A. (2017) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2017.8317848 · Cited by 15
Identification of Anomalous Behavioral Patterns in Crowd Scenes
Yang, T., Wang, C., Zhou, T., et al. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2022.022147 · Cited by 15
Trajectory distributions: A new description of movement for trajectory prediction
Lv, P., Wei, H., Gu, T., Zhang, Y., Jiang, X., Zhou, B., & Xu, M. (2021) · Computational Visual Media · doi:10.1007/s41095-021-0236-6 · Cited by 16
“We conducted experiments using the ETH dataset—see Table 3.”
BackgroundMethodInfluentialChen, Y., Liu, C., Mei, X., et al. (2022) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros47612.2022.9981037 · Cited by 15
“Baselines Ours Dataset SGAN Sophie Trajectron S-BiGAT CoMoGCN STGAT NMMP HGCN-GJS ETH 0.81/1.52 0.70/1.43 0.59/1.14 0.69/1.29 0.70/1.28 0.70/1.21 0.67/1.22 0.58/1.01 HOTEL 0.72/1.61 0.76/1.67 0.35/0.66 0.49/1.01 0.37/0.75 0.32/0.63 0.33/0.64 0.26/0.47 UNIV 0.60/1.26 0.54/1.24 0.54/1.13 0.55/1.32 0.53/1.16 0.56/1.20 0.…”
MethodInfluentialMotion instability based unsupervised online abnormal behaviors detection
Xie, S., & Guan, Y. (2015) · Multimedia Tools and Applications · doi:10.1007/s11042-015-2664-8 · Cited by 15
Abnormal Crowd Tracking and motion analysis
Santhiya, G., Sankaragomathi, K., Selvarani, S., et al. (2014) · IEEE International Conference on Advanced Communications, Control and Computing Technologies · doi:10.1109/icaccct.2014.7019309 · Cited by 15
“Recently, Mehran et al. [10] propose a method to model behaviours among a group of people.”
BackgroundMixture of Switching Linear Dynamics to Discover Behavior Patterns in Object Tracks
Kooij, J. F. P., Englebienne, G., & Gavrila, D. M. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2443801 · Cited by 15
“Probabilistic motion models are key to tracking [1], path prediction [5] in intelligent vehicles [2], and anomalous track detection in video surveillance [7].”
Unsupervised Pedestrian Trajectory Prediction with Graph Neural Networks
Wang, M., Shi, D., Guan, N., et al. (2019) · Proceedings - International Conference on Tools with Artificial Intelligence, TAI · doi:10.1109/ictai.2019.00119 · Cited by 15
A Sequential Monte Carlo Method for Multi-target Tracking with the Intensity Filter
Schikora, M., Koch, W., Streit, R. L., et al. (2012) · Studies in computational intelligence · doi:10.1007/978-3-642-28696-4_3 · Cited by 15
Sabeur, Z. A., Doulamis, N. D., Middleton, L., et al. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-27857-5_15 · Cited by 15
“The first class is to use techniques to isolate individuals, track them and estimate the crowd density and typical motions, while the second class models the crowd motion using fluid flow theory [9].”
Survey of Graph Neural Network for Internet of Things and NextG Networks
Moorthy, S. K., & Jagannath, J. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3625399 · Cited by 15
“In [179] the authors propose a GCN-based hybrid cloud-edge system called GPFS, a graph-based pose forecasting system that is capable of online learning that can significantly improve the forecasting accuracy in new environments.”
PSTG-based multi-label optimization for multi-target tracking
Chen, J., Sheng, H., Li, C., & Xiong, Z. (2016) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2015.06.002 · Cited by 14
Su, J., Huang, J., Qing, L., He, X., & Chen, H. (2022) · Heliyon · doi:10.1016/j.heliyon.2022.e11038 · Cited by 14
“introduced a Linear Trajectory Avoidance (LTA) model for walking people accounting for social interactions as well as scene knowledge with video data [7], whereas 3-D range sensors are exploited to person tracking in public spaces which are mounted above human height to have less occlusion between persons [8].”
BackgroundLiu, S., Chen, X., Wu, Z., et al. (2022) · Proceedings of the ... ACM International Conference on Information & Knowledge Management · doi:10.1145/3511808.3557345 · Cited by 14
“Two evaluation metrics are used to measure the prediction error (a lower metric stands for a better model performance): • Average Displacement Error (ADE): The Euclidean distance between the ground truth trajectories and the predicted trajectories averaged overall predicted time steps [31].”
MethodREGROUP: A Robot-Centric Group Detection and Tracking System
Taylor, A. M., & Riek, L. D. (2022) · ACM/IEEE International Conference on Human-Robot Interaction (HRI) · doi:10.1109/hri53351.2022.9889634 · Cited by 14
“They then employ probabilistic methods [29], [30], [31], [32], graph-based approaches [33], [34], [35], or social force models [36], [37], [38] to detect groups.”
MethodSang, H., Chen, W., Wang, H., et al. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15989-4 · Cited by 14
“The proposed method was validated in two classical datasets of ETH [18] and UCY [13], where ETH included two real scenarios (ETH and HOTEL), and UCY consisted of three real scenarios (ZARA1, ZARA2, and UNIV).”
Inverse Reinforcement Learning of Pedestrian–Robot Coordination
Gonon, D. J., & Billard, A. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3289770 · Cited by 14
“We assume that there is a scalar cost function J ( X , U ) which reflects the collective objectives of the multi-agent system under consideration, e.g. navigating efficiently and without any collision.”
Sun, J., Li, Y., Chai, L., & Lu, C. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.02108 · Cited by 14
Xu, Y., Jia, Q., Wang, H., et al. (2024) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2024.109361 · Cited by 14
Risk-Aware Vehicle Trajectory Prediction Under Safety-Critical Scenarios
Wang, Q., Xu, D., Kuang, G., Lv, C., Li, S. E., & Nie, B. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3525744 · Cited by 14
“Numerous trajectory prediction datasets have emerged to facilitate training algorithms [27], [28], [29].”
BackgroundEfficient crowd simulation in complex environment using deep reinforcement learning
Li, Y., Chen, Y., Liu, J., et al. (2025) · Scientific Reports · doi:10.1038/s41598-025-88897-2 · Cited by 14
Online Multi-Target Tracking With Unified Handling of Complex Scenarios
Jiang, H., Wang, J., Gong, Y., Rong, N., Chai, Z., & Zheng, N. (2015) · IEEE Transactions on Image Processing · doi:10.1109/tip.2015.2446331 · Cited by 13
“Since in real-world scenarios, people always interact with others and the environment, there also exist approaches that model social behaviors of pedestrians for more robust tracking, especially when occlusions exist [19], [20].”
BackgroundMethodCollaborative Tracking for Multiple Objects in the Presence of Inter-Occlusions
Xiao, J., & Oussalah, M. (2015) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2015.2406193 · Cited by 14
“The second employs the Linear Trajectory Avoidance (LTA) method proposed in [31], which has been proven to provide good accuracy results in case of multiple object tracking.”
MethodPredictive and Multirate Sensor-Based Planning Under Uncertainty
Mora, M. C., & Tornero, J. (2014) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2014.2366974 · Cited by 14
Tracking People and Their Objects
Baumgartner, T., Mitzel, D., & Leibe, B. (2013) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2013.469 · Cited by 14
A Multiple Targets Appearance Tracker Based on Object Interaction Models
Li, G., Qu, W., & Huang, Q. (2011) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2011.2165591 · Cited by 14
“First, in [13], a homography matrix is needed to transform image plane to world coordinates which requires users manually click points.”
BackgroundMethodResultInfluentialState estimation for on–off nonlinear stochastic coupling networks with time delay
Li, W., Jia, Y., & Du, J. (2016) · Neurocomputing · doi:10.1016/j.neucom.2016.09.012 · Cited by 14
A nearly optimal landmark deployment for indoor localisation with limited sensing
Magnago, V., Palopoli, L., Passerone, R., et al. (2017) · International Conference on Indoor Positioning and Indoor Navigation · doi:10.1109/ipin.2017.8115883 · Cited by 14
“performance of landmark greedy placement [27].”
BaG: Behavior-aware Group Detection in Crowded Urban Spaces using WiFi Probes
Shen, J., Cao, J., & Liu, X. (2019) · International World Wide Web Conference (WWW) · doi:10.1145/3308558.3313590 · Cited by 14
“The grouping information is an essential contextual feature in multi-target tracking [6] and behavior analysis [7].”
BackgroundDetecting and Tracking of Multiple Mice Using Part Proposal Networks
Jiang, Z., Liu, Z., Chen, L., Tong, L., Zhang, X., Lan, X., Crookes, D., Yang, M., & Zhou, H. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3160800 · Cited by 17
“To improve the multi-target tracking accuracy, several methods ([53]–[55]) have also presented group models, in which each object is considered of having the relationship with other objects and surroundings.”
MethodDeep Context Maps: Agent Trajectory Prediction Using Location-Specific Latent Maps
Gilitschenski, I., Rosman, G., Gupta, A., Karaman, S., & Rus, D. L. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3004800 · Cited by 13
“ETH Dataset [40] The ETH dataset, also known as BIWI Walking Pedestrian dataset, is a collection of two sequences (ETH and HOTEL) recorded around ETH Zurich.”
MethodEstimating Pedestrian Crossing States Based on Single 2D Body Pose
Wang, Z., & Papanikolopoulos, N. P. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341745 · Cited by 14
“Previous prediction-based research computes the pedestri-ans’ incoming moving trajectory or predicts their crossing intention based on previous trajectories [2] or velocities [3].”
MethodModeling human–human interaction with attention-based high-order GCN for trajectory prediction
Fang, Y., Jin, Z., Cui, Z., et al. (2021) · The Visual Computer · doi:10.1007/s00371-021-02109-2 · Cited by 14
“The ETH dataset & the UCY dataset contains five crowd scenes.”
BackgroundMethodInfluentialSTUGCN:A Social Spatio-Temporal Unifying Graph Convolutional Network for Trajectory Prediction
Zhao, Z., & Liu, C. (2021) · International Conference on Automation, Control and Robotics Engineering (CACRE) · doi:10.1109/cacre52464.2021.9501325 · Cited by 14
The Stanford Drone Dataset Is More Complex Than We Think: An Analysis of Key Characteristics
Andle, J., Soucy, N., Socolow, S., et al. (2022) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2022.3166642 · Cited by 14
“Two datasets which have previously been used as benchmarks are ETH Zurich’s BIWI Walking Pedestrian Dataset [4] and UCY’s Crowds by Example [5].”
MethodIntegrated multiple behavior models for abnormal crowd behavior detection
Cho, S., & Kang, H.-B. (2012) · IEEE Southwest Symposium on Image Analysis and Interpretation (SSIAI) · doi:10.1109/ssiai.2012.6202466 · Cited by 14
“On the other hand, social factor based method[3-6] has its own drawback; it has difficulties to detect the personal property such as a person’s abnormal movement.”
BackgroundPedestrian trajectory prediction using BiRNN encoder–decoder framework
Wu, J., Woo, H., Tamura, Y., et al. (2019) · Advanced Robotics · doi:10.1080/01691864.2019.1635910 · Cited by 14
“We also evaluated the percentage of the successful prediction with ADE smaller than 1 m to compare with earlier machine learning based approach [18] which is similar to those model-based approaches since it only has a few trainable parameters.”
BackgroundMethodInfluentialOnline State-Time Trajectory Planning Using Timed-ESDF in Highly Dynamic Environments
Zhu, D., Zhou, T., Lin, J., Fang, Y., & Meng, M. Q. ‐. (2022) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra46639.2022.9812436 · Cited by 14
“(b) Benchmark datasets [34], [35].”
Metadata
- Authors
- Pellegrini, S., Ess, A., Schindler, K., & van Gool, L.
- Affiliations
- ETH Zurich (CH)
- Technische Universität Darmstadt (DE)
- Venue
- 2009 IEEE 12th International Conference on Computer Vision
- Publisher
- IEEE
- Published
- 2009-09
- Pages
- 261-268
- OpenAlex
- W2532516272
- Semantic Scholar
- 7d7a98a81715e2d3747071722b2deeed8937d122
- Cited by
- 1,683
- References
- 25
- Citations by year
- 2026: 85 · 2025: 185 · 2024: 192 · 2023: 163 · 2022: 153 · 2021: 156 · 2020: 141 · 2019: 100 · 2018: 66 · 2017: 68 · 2016: 86 · 2015: 69 · 2014: 75 · 2013: 54 · 2012: 42
- FWCI
- 24.50
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- IEEE/CVF International Conference on Computer Vision (ICCV) (publishedVersion)
- Lirias (KU Leuven) (submittedVersion)
- Funders
- Else Kröner-Fresenius-Stiftung
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Surveillance and Tracking Methods
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Social Sciences › Safety Research › Evacuation and Crowd Dynamics
- Keywords
- multi-target tracking, social behavior modeling, crowd simulation, collision avoidance, occlusion handling, vehicle-mounted camera, bird's-eye view, data association, human trajectory prediction, social interaction
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
