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 801–900 of 1,000
The Group and Crowd Analysis Interdisciplinary Challenge
Murino, V., Cristani, M., Shah, S. K., et al. (2017) · Elsevier eBooks · doi:10.1016/b978-0-12-809276-7.00001-1 · Cited by 10
Recent advances in human behaviour understanding: A survey
Mahata, J., & Phadikar, A. (2017) · Devices for Integrated Circuit (DevIC) · doi:10.1109/devic.2017.8074052 · Cited by 10
“Some algorithms were proposed to detect multiple objects [12] which can be used to develop a system for detecting multiple persons in a scene and their interactions.”
BackgroundWirthmüller, F., Klimke, M., Schlechtriemen, J., Hipp, J., & Reichert, M. (2020) · IEEE Symposium Series on Computational Intelligence (SSCI) · doi:10.1109/ssci47803.2020.9308542 · Cited by 10
Ensemble of LSTMs and Feature Selection for Human Action Prediction
Petković, T., Petrović, L., Marković, I., & Petrovic, I. K. (2022) · Lecture notes in networks and systems · doi:10.1007/978-3-030-95892-3_33 · Cited by 10
“Examples of purely motion prediction datasets are: KITTI [7], ETH [17] and UCY [15].”
MethodConditional Generative Adversarial Networks for Speed Control in Trajectory Simulation
Julka, S., Sowrirajan, V., Schloetterer, J., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-95470-3_33 · Cited by 10
“As shown in [42], these datasets also cover challenging group behaviours such as couples walking together, groups crossing each other and groups forming and dispersing in some scenes, and contain several other non-linear trajectories.”
BackgroundSocial-DPF: Socially Acceptable Distribution Prediction of Futures
Shi, X., Shao, X., Wu, G., et al. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i3.16357 · Cited by 10
“The two datasets contain 5 sets, which are UCY-zara01, UCY-zara02, UCY-univ, ETH-hotel, ETH-eth in 4 crowded scenarios with totally 1536 trajectories.”
MethodInfluentialPedestrian Trajectory Prediction with MLP-social-GRU
Zhang, Y., & Zheng, L. (2021) · International Conference on Machine Learning and Computing (ICMLC) · doi:10.1145/3457682.3457737 · Cited by 10
“In this section, we evaluate our predictor on two public datasets, ETH [21] and UCY [22].”
MethodOptimization-based Pedestrian Model Calibration for Evaluation
Wolinski, D., Guy, S. J., Olivier, A., et al. (2014) · Transportation research procedia · doi:10.1016/j.trpro.2014.09.041 · Cited by 10
Representing dense crowd patterns using bag of trajectory graphs
Zhang, Y., Huang, Q., Qin, L., et al. (2014) · Signal Image and Video Processing · doi:10.1007/s11760-014-0669-9 · Cited by 10
“[24] consider moving entities as agents and model them using social behavior analysis.”
BackgroundMethodCrowd Tracking by Group Structure Evolution
Zhu, F., Wang, X., & Yu, N. (2016) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2016.2615460 · Cited by 10
“Recently, the social force model [35] was used to track multiple pedestrians by modeling their interactions, the influ-ence of destinations, and scene structures, with a physical model [36], [37].”
BackgroundMethodAn RNN-Based IMM Filter Surrogate
Becker, S., Hug, R., Hübner, W., & Arens, M. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-20205-7_32 · Cited by 11
“s for stopping examples in which the maximum of the predicted distribution is visible close to the last observation. The performance is compared with the nal displacement error (FDE) (see for example [22]) of the lateral motion (from the vehicle perspective) for three dierent time horizons, in particular 8 steps (0:5s…”
Pedestrian Occupancy Prediction for Autonomous Vehicles
Zechel, P., Streiter, R., Bogenberger, K., et al. (2019) · IEEE International Conference on Robotic Computing (IRC) · doi:10.1109/irc.2019.00042 · Cited by 10
“Pellegrini [4] presents a model in the shape of Linear Trajectory Avoidance (LTA) which analyzes the interactions with other people not just on the basis of the current position, but extrapolates their behavior.”
BackgroundReal Time Crowd Navigation from First Principles of Probability Theory
Trautman, P., & Patel, K. (2020) · Proceedings of the International Conference on Automated Planning and Scheduling · doi:10.1609/icaps.v30i1.6741 · Cited by 10
“Alternatively, many trajectories in the ETH and UCY datasets little or no can tell that ETH and UCY is mostly linear by looking at the results in recent studies that benchmarked over the whole of both datasets (Alahi, Goel, and et al 2016), (Gupta et al. 2018): linear extrapolation error rates were very low (e.g., 0.3…”
MethodInfluential3D Pedestrian tracking using local structure constraints
Nguyen, U., & Heipke, C. (2020) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2020.05.002 · Cited by 11
Frame-wise Motion and Appearance for Real-time Multiple Object TrackingPreprint
Zhang, J., Zhou, S., Wang, J., & Huang, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1905.02292 · Cited by 10
“In earlier tracking approaches, the color histogram [9, 24, 26, 49] and pixel-based template representations [35, 53] are standard hand-crafted features used to describe the appearance of objects.”
MethodGhiță, A. Ș., & Florea, A. M. (2022) · Sensors · doi:10.3390/s22155856 · Cited by 9
“In particular, for the ETH dataset the results of the segmentation technique are not as good as in the case of the UCY dataset.”
InfluentialAutofocusing for Synthetic Aperture Imaging Based on Pedestrian Trajectory Prediction
Pei, Z., Zhang, J., Zhang, W., et al. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3314895 · Cited by 9
“Datasets [14], [15] which are used by many existing trajectory prediction methods are acquired by a single camera.”
BackgroundCSCMOT: Multi-object tracking based on channel spatial cooperative attention mechanism
Wang, F., Yan, H., Zhang, L., & Gao, K. (2023) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107084 · Cited by 9
Learning Relation in Crowd Using Gated Graph Convolutional Networks for DRL-Based Robot Navigation
Jiang, H., Bhujel, N., Lin, Z., Wan, K.-W., Li, J., Senthilnath, J., & Jiang, X. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3343923 · Cited by 9
“ETH dataset [45] includes the HOTEL ( ∼ 6 pedestrians/frame) and ETH dataset ( ∼ 6 pedes-trians/frame).”
MethodA Comprehensive Review on Group Re-identification in Surveillance Videos
Nayak, K., & Dogra, D. P. (2025) · ACM Computing Surveys · doi:10.1145/3711126 · Cited by 11
“The approach has been tested on the BIWI dataset [ 81 ] and the Crowd by Example dataset [ 62 ].”
Low-FPS Multi-Object Multi-Camera Tracking via Deep Learning
Hsieh, Y.-H., Kao, C.-C., Lai, C.-H., Lin, K., Yang, S.-Y., & Yuan, S. (2025) · Electronics · doi:10.3390/electronics14071373 · Cited by 10
SIAT: Pedestrian trajectory prediction via social interaction-aware transformer
Wang, C., Wang, J., Gao, W., et al. (2025) · Complex & Intelligent Systems · doi:10.1007/s40747-025-01944-3 · Cited by 9
“Extensive experiments demonstrate that the proposed SIAT significantly improves both accuracy and robustness on pedestrian trajectory prediction datasets,including ETH-UCY [16, 17] and SDD [18].”
Fan, K., Dong, W., Liu, H., et al. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3589759 · Cited by 9
“We evaluate our proposed model on the ETH/UCY [43] dataset and Stanford Drone Dataset (SDD) [44], which are widely used for multi-agent trajectory prediction tasks.”
Social grouping for target handover in multi-view video
Qin, Z., Shelton, C. R., & Chai, L. (2013) · IEEE International Conference on Multimedia and Expo (ICME) · doi:10.1109/icme.2013.6607433 · Cited by 9
“Researchers have used various social factors to improve tracking performance, including a pedestrian’s destination, desired speed, repulsion from other individuals, and social grouping behavior [15, 16, 17, 18].”
BackgroundMultimodal Pedestrian Trajectory Prediction Based on Relative Interactive Spatial-Temporal Graph
Zhao, D., Li, T., Zou, X., He, Y., Zhao, L., Chen, H., & Zhuo, M. (2022) · IEEE Access · doi:10.1109/access.2022.3200066 · Cited by 9
“We evaluate the model performance on two common pedestrian trajectory datasets: ETH [47] and UCY [48].”
MethodRobot navigation in crowds via deep reinforcement learning with modeling of obstacle uni-action
Lu, X., Woo, H., Faragasso, A., et al. (2022) · Advanced Robotics · doi:10.1080/01691864.2022.2142068 · Cited by 9
“Aside from the simulation experiments above, we also trained CADRL, SARL, SOADRL, and our methods, OUDRL, with the ETH dataset [33] and then implemented the learned policy to navigate the robot in the Hotel dataset [33] and theNYCGrand Central (NYC-GC) dataset [34].”
MethodEntry-Flipped Transformer for Inference and Prediction of Participant Behavior
Hu, B., & Cham, T. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19772-7_26 · Cited by 9
Disentangling Crowd Interactions for Pedestrians Trajectory Prediction
Bhujel, N., & Yau, W. (2023) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2023.3262420 · Cited by 9
“The ETH [11] dataset consists of two scenes; ETH and HOTEL, whereas the UCY [2] dataset consists of three scenes; ZARA1, ZARA2, and UNIV.”
BackgroundMethodPerspective Distortion Model for Pedestrian Trajectory Prediction for Consumer Applications
Gundreddy, S., Ramkumar, R., Raman, R., et al. (2023) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2023.3318050 · Cited by 9
“To predict short-term trajectory, a linear trajectory avoidance model was introduced by Pellegrini et al. [14].”
MethodGigaTraj: Predicting Long-term Trajectories of Hundreds of Pedestrians in Gigapixel Complex Scenes
Lin, H., Wei, C., He, L., Guo, Y., Zhao, Y., Li, S., & Fang, L. (2024) · 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/cvpr52733.2024.01829 · Cited by 10
“On the one hand, datasets with narrow field of view (FoV), such as ETH [1]/UCY [2], cannot observe long-term trajectories.”
BackgroundMethodInfluentialModeling crowd motions for abnormal activity detection
Lee, D.-G., Suk, H., & Lee, S.-W. (2014) · IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) · doi:10.1109/avss.2014.6918689 · Cited by 9
“The interaction energy potential was estimated from the velocity of space-time interest points to explain whether they would meet in near future [19].”
BackgroundCrowd Behavior Recognition Using Dense Trajectories
Khokher, M. R., Bouzerdoum, A., & Phung, S. L. (2014) · International Conference on Digital Image Computing: Techniques and Applications (DICTA) · doi:10.1109/dicta.2014.7008098 · Cited by 9
Pedestrian Trajectory Prediction Using a Social Pyramid
Xue, H., Huynh, D. Q., & Reynolds, M. A. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-29911-8_34 · Cited by 10
Multiple human trajectory prediction and cooperative navigation modeling in crowded scenes
Hacınecipoğlu, A., Konukseven, E. İ., & Koku, A. B. (2020) · Intelligent Service Robotics · doi:10.1007/s11370-020-00333-8 · Cited by 9
“To test the proposed method, we use ETH BIWI Walking Pedestrians dataset [9].”
BackgroundMethodInfluentialA Graphical Social Topology Model for RGB-D Multi-Person Tracking
Gao, S., Ye, Q., Liu, L., Kuijper, A., & Ji, X. (2021) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2021.3049397 · Cited by 10
“[19] proposed an effective dynamic group model, considering nearby pedestri-ans’ positions.”
BackgroundMethodBEV-Net: Assessing Social Distancing Compliance by Joint People Localization and Geometric Reasoning
Dai, Z., Jiang, Y., Li, Y., Liu, B., Chan, A. B., & Vasconcelos, N. M. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00535 · Cited by 9
“One line of work achieves this through the task of trajectory prediction [34, 49, 29, 1, 19], which requires generating plausible motion paths for pedestrians in the image plane.”
BackgroundSocial-Transformer: Pedestrian Trajectory Prediction in Autonomous Driving Scenes
Sun, H., & Sun, F. (2022) · Communications in computer and information science · doi:10.1007/978-981-16-9247-5_13 · Cited by 9
The utility of datasets in crowd modelling and analysis: a survey
Waqar, S., Khan, M. U. G., Waseem, M. H., & Qayyum, S. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13227-x · Cited by 9
“This dataset has been used for people tracking [46].”
BackgroundMethodInfluentialPetković, T., Petrović, L., Marković, I., & Petrovic, I. K. (2022) · Applied Soft Computing · doi:10.1016/j.asoc.2022.109245 · Cited by 9
End-to-end video compression for surveillance and conference videos
Wang, S., Zhao, Y. Q., Gao, H., Ye, M., & Li, S. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13484-w · Cited by 10
“Since our network aims at surveillance and conference videos, the test sets we use are HEVC standard test sequences (Class A, Class E) [28], CUHK Square surveillance dataset [33] and Ewap surveillance dataset (Ewap eth and Ewap hotel) [21].”
MethodAdaptive trajectory prediction without catastrophic forgetting
Zhi, C., Sun, H., & Xu, T. (2023) · The Journal of Supercomputing · doi:10.1007/s11227-023-05241-z · Cited by 9
“[18] Quantitatively and objectively evaluated the potential domain differences between the ETH and UCY experimental datasets.”
MethodInfluentialA novel dynamic model for multiple pedestrians tracking in extremely crowded scenarios
Song, X., Shao, X., Zhang, Q., Shibasaki, R., Zhao, H., & Zha, H. (2012) · Information Fusion · doi:10.1016/j.inffus.2012.08.004 · Cited by 9
Multi-object tracking via high accuracy optical flowand finite set statistics
Schikora, M., Koch, W., & Cremers, D. (2011) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2011.5946677 · Cited by 9
Group Detection and Tracking Using Sociological Features
Vascon, S., & Bazzani, L. (2017) · Elsevier eBooks · doi:10.1016/b978-0-12-809276-7.00004-7 · Cited by 9
Instant coherent group motion filtering by group motion representations
Li, N., Zhang, Y., Luo, W., et al. (2017) · Neurocomputing · doi:10.1016/j.neucom.2017.05.045 · Cited by 9
Anomaly Detection Using Motion Patterns Computed from Optical Flow
Parvathy, R., Thilakan, S., Joy, M., et al. (2013) · International Conference on Advances in Computing and Communications (ICACC) · doi:10.1109/icacc.2013.18 · Cited by 9
“Another method to detect and normalize abnormal behavior is using the social force model.”
MethodBrowarek, S. (2010) · DSpace@MIT (Massachusetts Institute of Technology) · Cited by 9
Real-time Crowd Tracking using Parameter Optimized Mixture of Motion ModelsPreprint
Bera, A., Wolinski, D., Pettré, J., et al. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1409.4481 · Cited by 9
International Journal of Ethics in Engineering & Management Education
vision-i, P. (2014) · ADVANCED SCIENCES INDEX · Cited by 9
Modeling and Inferring Human Intents and Latent Functional Objects for Trajectory PredictionPreprint
Xie, D., Shu, T., Todorović, S., et al. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1606.07827 · Cited by 9
“People’s trajectories have also been estimated based on inferring social interactions [28], [29], [30], [31], or detecting objects in egocentric videos [32].”
BackgroundKnowledge Transfer for Scene-specific Motion PredictionPreprint
Ballan, L., Castaldo, F., Alahi, A., Palmieri, F., & Savarese, S. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1603.06987 · Cited by 9
“To this end, social forces and contextual relationships have been effectively used to model human-human interactions in crowded environments [23,27,43,1,20,3].”
Potential Field: Interpretable and Unified Representation for Trajectory PredictionPreprint
Su, S., Peng, C., Shi, J., & Choi, C. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1911.07414 · Cited by 10
“We evaluate our algorithm on three widely used benchmark datasets ETH [22] / UCY [20] and Stanford Drone Dataset (SDD) [24].”
MethodDufour, O., Dang, H.-T., Cordes, J., et al. (2025) · Scientific Data · doi:10.1038/s41597-025-04732-3 · Cited by 9
“For instance, the widely used ETH 14 and UCY 15 datasets, which originate from surveillance videos, capture pedestrian scenes at density 0.1-0.5 ped/m 2 , with many avoidance situations.”
InfluentialPedestrian trajectory prediction method based on the Social-LSTM model for vehicle collision
Han, Y. O., Lin, X., Pan, D. L., et al. (2023) · Transportation Safety and Environment · doi:10.1093/tse/tdad044 · Cited by 8
Pedestrian Crossing Direction Prediction at Intersections for Pedestrian Safety
Kim, Y., Abdel-Aty, M. A., Choi, K., et al. (2025) · IEEE Open Journal of Intelligent Transportation Systems · doi:10.1109/ojits.2025.3574082 · Cited by 8
“…for pedestrian intention prediction. c: Dataset for Pedestrian Behavior Analysis: The availability of extensive open datasets such as PIE [32], JAAD [33], SDD [36], ETH [34], and UCY [35] have significantly facilitated research on pedestrian intention prediction in streets within urban areas.”
Xu, F., Zhang, B., & Liu, Y. (2026) · Electronics · doi:10.3390/electronics15081738 · Cited by 8
Xiao, J., Stolkin, R., & Leonardis, A. (2014) · British Machine Vision Conference (BMVC) · doi:10.5244/c.28.101 · Cited by 8
“Fortunately, the latent team strategy can provide additional hints for tracking, i.e. seemingly erratic motion of a single player may actually be consistent with the overall motion strategy of the team.”
BackgroundObstacle‐transformer: A trajectory prediction network based on surrounding trajectories
Zhang, W., Chai, Q., Zhang, Q., & Wu, C. (2022) · IET Cyber-Systems and Robotics · doi:10.1049/csy2.12066 · Cited by 10
“TABLE 2 Quantitative results of different algorithms on ETH and UCY datasets Metric Dataset Linear LSTM S‐GAN [22] S‐LSTM [9] Next [31] TF [8] ST‐Att.”
MethodInfluentialReciprocal Consistency Prediction Network for Multi-Step Human Trajectory Prediction
Zhu, W., Liu, Y., Zhang, M., et al. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3252262 · Cited by 8
“We follow the standard test train split as used in [34].”
BackgroundMethodA Dual Perspective of Human Motion Analysis - 3D Pose Estimation and 2D Trajectory Prediction
Zaier, M., Wannous, H., Drira, H., et al. (2023) · IEEE/CVF International Conference on Computer Vision Workshops (ICCVW) · doi:10.1109/iccvw60793.2023.00233 · Cited by 8
“Regarding 2D trajectory prediction, our approach was evaluated on well-established public human-trajectory datasets, namely ETH [37] and UCY [23] datasets, which are widely-used benchmarks for pedestrian motion prediction.”
MethodA Study in Zucker: Insights on Interactions Between Humans and Small Service Robots
Day, A., & Karamouzas, I. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3355641 · Cited by 8
“In benchmarks commonly used for human trajectory prediction, such as ETH [35], UCY [47], and SDD [48], it has been shown that CVM-based approaches can rival even state-of-the-art neural network approaches [49].”
MethodDeep Learning Based Trajectory Prediction in Autonomous Driving Tasks: A Survey
Xing, H., Liu, W., Ning, Z., Zhao, Q., Cheng, S., & da Hu, J. (2024) · International Conference on Computer and Automation Engineering · doi:10.1109/iccae59995.2024.10569771 · Cited by 8
“Finally, we compare the performance of these trajectory prediction models on public datasets [1] [5] [6] and discuss potential directions for future research.”
Robustness in trajectory prediction for autonomous vehicles: a survey
Hagenus, J., Mathiesen, F. B., Schumann, J., & Zgonnikov, A. (2024) · 2024 IEEE Intelligent Vehicles Symposium (IV) · doi:10.1109/iv55156.2024.10588389 · Cited by 7
“For this a number of measures have been developed [20]–[25].”
BackgroundMethodTrACT: A Training Dynamics Aware Contrastive Learning Framework for Long-Tail Trajectory Prediction
Zhang, J., Pourkeshavarz, M., & Rasouli, A. (2024) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv55156.2024.10588635 · Cited by 8
“We evaluate TrACT on two benchmark trajectory prediction datasets: nuScenes [3] and ETH-UCY [10], [11]. nuScenes is a large-scale autonomous driving dataset containing 1000 scenes, which are split into 850 and 150 scenes for training-validation and testing, respectively.”
MethodBarua, H. B., Mg, T. H., Pramanick, P., & Sarkar, C. (2024) · ACM Transactions on Human-Robot Interaction · doi:10.1145/3682072 · Cited by 15
Multi-modal Knowledge Distillation-based Human Trajectory Forecasting
Jeong, J., Lee, S., Park, D.-H., Lee, G., & Yoon, K. (2025) · 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/cvpr52734.2025.02256 · Cited by 11
“Our framework is validated on three pedestrian trajectory forecasting datasets (JRDB [39], SIT [4], and ETH[47]/UCY [27]) and with two state-of-the-art trajectory forecasting models (HiVT [69] and MART [26]).”
Tracking with a mixed continuous-discrete Conditional Random Field
Pellegrini, S., & Gool, L. V. (2012) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2012.09.005 · Cited by 8
“These findings have found applications in several disciplines, among which we mention computer graphics [43,40], robotics [57,22] and computer vision, where they have also been employed for tracking [2,41,51].”
BackgroundMethodSelf-citationUnderstanding dyadic interactions applying proxemic theory on videosurveillance trajectories
Calderara, S., & Cucchiara, R. (2012) · 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops · doi:10.1109/cvprw.2012.6239351 · Cited by 8
“Basically, social theories of space perception have been applied to improve tracking in crowd using Helbing’s social force model [9], have been injected in tracking models to update predictions, [14], and have been studied in social robotics science in order to improve interaction between human and mobile robots [2].”
MethodOnline multiperson tracking with occlusion reasoning and unsupervised track motion model
McLaughlin, N., del Rincón, J. M., & Miller, P. C. (2013) · IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS) · doi:10.1109/avss.2013.6636613 · Cited by 8
“1 we can observe that the TA results from our approach are better than, or comparable with the literature, even against techniques based on global optimisation [24, 9], or including more complex reasoning, such as social behaviour [13, 18, 23].”
ResultA video surveillance system designed to detect multiple falls
Chen, M.-C. (2016) · Advances in Mechanical Engineering · doi:10.1177/1687814016642914 · Cited by 8
Future Near-Collision Prediction from Monocular Video: Feasibility, Dataset, and Challenges.
Manglik, A., Weng, X., Ohn-Bar, E., & Kitani, K. (2019) · Cited by 11
Attention Modulated Multiple Object Tracking with Motion Enhancement and Dual Correlation
Wang, Y., Zhang, Z., Zhang, N., & Zeng, D. (2021) · Symmetry · doi:10.3390/sym13020266 · Cited by 8
“In this two-stage approach, earlier literature works use color histogram [10–12] and pixel-based template representation [13,14] features to describe the objects.”
MethodMotion Prediction in Visual Object Tracking
Wang, J., & He, Y. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341158 · Cited by 8
“In the early days, motion model was a core component of tracking - constant velocity models [2], Kalman filters [3], particle filters [4] and even social force models [5] for more complicated motions.”
MethodBecker, S., Hug, R., Huebner, W., et al. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-90439-5_41 · Cited by 8
“For a comparison to other approaches, the publicly available BIWI [27] and UCY [22] datasets are used according to the common practice of fully observed input data.”
MethodInfluentialReconstruction of fragmented trajectories of collective motion using Hadamard deep autoencoders
Gajamannage, K., Park, Y., Paffenroth, R. C., et al. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.108891 · Cited by 8
“With the advancements in both the video recording technology and the multi-object tracking (MOT) methods, dynamics of animal groups in wild [1], pedestrians on streets [2], vehicles on roads [3], and sport players on courts [4] are studied using agents’ individual trajectories.”
Multi-object tracking based on network flow model and ORB feature
Chen, J., Xi, Z., Lu, J., & Ji, J. (2022) · Applied Intelligence · doi:10.1007/s10489-021-03042-6 · Cited by 8
Towards Benchmarking Human-Aware Social Robot Navigation: A New Perspective and Metrics
Singamaneni, P. T., Favier, A., & Alami, R. (2023) · IEEE RO-MAN · doi:10.1109/ro-man57019.2023.10309398 · Cited by 8
“The similarity in trajectory is quantified using metrics that measure the path deviation like Average Displacement Error (ADE) [14], Final Displacement Error (FDE) [15] and their variations.”
MethodLi, K., Chen, Y., Shan, M., Li, J., Worrall, S., & Nebot, E. (2023) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc57777.2023.10422696 · Cited by 8
“In this work, the Vehicle-Crowd Interaction (VCI), ETH [39] and UCY [40] are utilized.”
MethodMethods to model the motion of extended objects in multi-object Bayes filters
Reuter, S., Wilking, B., & Dietmayer, K. (2012) · Cited by 8
Co-operative Pedestrians Group Tracking in Crowded Scenes Using an MST Approach
Setia, A., & Mittal, A. (2015) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2015.21 · Cited by 8
Activity recognition and prediction with pose based discriminative patch model
Cao, S., Chen, K., & Nevatia, R. (2016) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2016.7477584 · Cited by 8
“Most of the works focused on predicting (or forecasting) the future trajectories of pedestrians [11][18].”
BackgroundFast Object Motion Estimation Based on Dynamic Stixels
Morales, N., Morell, A., Toledo, J. T., et al. (2016) · Sensors · doi:10.3390/s16081182 · Cited by 8
Park, S.-H., Lee, K., & Yoon, K. (2016) · Proceedings - International Conference on Image Processing · doi:10.1109/icip.2016.7533007 · Cited by 7
“For robust MOT under moving camera scenes, sophisticated motion models such as social force models [9] [10] [11] and relative motion models [12] [13] are proposed, too.”
MethodGabbard, R. D., Fendley, M. E., Dar, I. A., Warren, R., & Kashou, N. H. (2017) · Neurophotonics · doi:10.1117/1.nph.4.4.041406 · Cited by 8
Recurrent Neural Network Architectures for Vulnerable Road User Trajectory Prediction
Xiong, H., Flohr, F. B., Wang, S., Wang, B., Wang, J., & Li, K. (2019) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2019.8814275 · Cited by 8
“Commonly used metrics for the evaluation of motion trajectory prediction is the Average Displacement Error [34] and the Final Displacement Error of the ground point [5], [21] or the bottom center of bounding box in meters [18], others include the center location error of the bounding box (computed as mean squared erro…”
MethodVelocity-Based Models for Crowd Simulation
Pettré, J., Wolinski, D., & Olivier, A. (2013) · Springer eBooks · doi:10.1007/978-3-319-02447-9_88 · Cited by 8
SALSA: A Novel Dataset for Multimodal Group Behavior AnalysisPreprint
Alameda-Pineda, X., Staiano, J., Subramanian, R., et al. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1506.06882 · Cited by 8
“improve by incorporating aspects of human social behavior [19], [20], as well as with activity analysis [21].”
BackgroundRealtime Multilevel Crowd Tracking using Reciprocal Velocity ObstaclesPreprint
Bera, A., & Manocha, D. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1402.2826 · Cited by 8
“Other motion models used in pedestrian tracking are the Social Force model [28] [29] [3], LTA [30], RVO [31] and ATTR [32].”
MethodLe, V.-A., Tadiparthi, V., Chalaki, B., et al. (2024) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra57147.2024.10611204 · Cited by 7
“Trajnet++ benchmark for training Social-LSTM models on the ETH dataset [32].”
BackgroundDual-Alignment Domain Adaptation for Pedestrian Trajectory Prediction
Li, W., Li, F., Jing, X., et al. (2024) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2024.3481831 · Cited by 7
“We evaluate the performance of our method on two widely used public datasets over previous methods: ETH [35] and UCY [36].”
MethodGroup-PTP: A Pedestrian Trajectory Prediction Method Based on Group Features
Zhang, C., Zhang, G., Zheng, Z., et al. (2025) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2025.3535380 · Cited by 7
“Previous studies [4], [5], [6], [17], [18] relied on rule-based methods that constrain interactions to local scopes, thereby making it challenging to simulate complex group interactions.”
Chib, P. S., & Singh, P. P. (2025) · Information Fusion · doi:10.1016/j.inffus.2025.103262 · Cited by 10
Zhang, H., Shi, L., Shi, L., Liu, C., Gao, Z., Xu, Y., & Chen, L. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.130464 · Cited by 7
Timewise Intentions and Time-Varying Distribution Network for Pedestrian Trajectory Prediction
Wang, R., Cheng, J., & Yu, J. (2025) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2025.3594563 · Cited by 7
“To demonstrate the prediction performance of our model, we perform the experiments on Joint Attention for Autonomous Driving (JAAD) [39], Pedestrian Intention Estimation (PIE) [40], ETH [41] and UCY [42], and Stanford Drone Dataset (SDD) [43].”
Yang, F., Saikia, H., & Peters, C. (2018) · doi:10.1145/3267851.3267875 · Cited by 7
“We set ETH and UCY as low-density level, and PWP as high-density level.”
MethodInfluentialTrajectory Prediction from Hierarchical Perspective
Qian, T., Xu, Y., Zhang, Z., et al. (2022) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3503161.3548092 · Cited by 7
“In the medium and right columns (i.e, hotel, and ETH scenes), we observe that there is a slight difference between ground truth and predicted trajectory.”
MethodDiscovering and Understanding Algorithmic Biases in Autonomous Pedestrian Trajectory Predictions
Bae, A., & Xu, S. (2022) · ACM Conference on Embedded Networked Sensor Systems (SenSys) · doi:10.1145/3560905.3568433 · Cited by 8
“First off, we were unable to evaluate on other popular trajectory prediction datasets, such as ETH [25], UCY [19], or nuScenes [5], due to their lack of pedestrian demographic labels.”
MethodAVstack: An Open-Source, Reconfigurable Platform for Autonomous Vehicle Development
Hallyburton, R. S., Zhang, S., & Pajić, M. (2023) · ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS) · doi:10.1145/3576841.3585930 · Cited by 7
“Also, prediction error, ADE and FDE, are lower with collaboration compared to C1-base.”
InfluentialSocial Lode: Human Trajectory Prediction with Latent Odes
Ke, K., Yang, J., Liu, Y., et al. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10446317 · Cited by 7
Korbmacher, R., & Tordeux, A. (2024) · Sensors · doi:10.3390/s24072356 · Cited by 8
“Famous field datasets are the ETH [29] and UCY [30] datasets, which are widely used in the machine learning community.”
BackgroundMethod
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
