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 101–200 of 1,000
SALSA: A Novel Dataset for Multimodal Group Behavior Analysis
Alameda-Pineda, X., Staiano, J., Subramanian, R., Batrinca, L., Ricci, E., Lepri, B., Lanz, O., & Sebe, N. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2496269 · Cited by 171
“improve by incorporating aspects of human social behavior [19], [20], as well as with activity analysis [21].”
BackgroundIntent-aware long-term prediction of pedestrian motion
Karasev, V. I., Ayvacı, A., Heisele, B., & Soatto, S. (2016) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2016.7487409 · Cited by 157
“A number of models of “social” or “socially compliant” behavior have been proposed [6], [7], [8], [9], [10], [11].”
BackgroundMulti-camera multi-object tracking: A review of current trends and future advances
Amosa, T. I., Sebastian, P., Izhar, L. I., Ibrahim, O., Ayinla, S. L., Bahashwan, A. A., Bala, A., & Samaila, Y. A. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126558 · Cited by 159
Measuring Crowd Collectiveness
Zhou, B., Tang, X., Zhang, H., & Wang, X. (2014) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2014.2300484 · Cited by 151
EqMotion: Equivariant Multi-Agent Motion Prediction with Invariant Interaction Reasoning
Xu, C., Tan, R. T., Tan, Y.-H., Chen, S., Wang, Y. G., Wang, X., & Wang, Y. (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/cvpr52729.2023.00142 · Cited by 150
“ETH-UCY dataset [7, 10], contains 5 subsets, ETH, HOTEL, UNIV, ZARA1, and ZARA2.”
Forecasting Interactive Dynamics of Pedestrians with Fictitious Play
Ma, W.-C., Huang, D.-A., Lee, N., & Kitani, K. (2017) · 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.2017.493 · Cited by 148
“NLL nMDP[21] MDPCV mTA[30] FP FP + Speed Zara [25] 46.”
BackgroundMethodInfluentialRobot Navigation in Crowds by Graph Convolutional Networks With Attention Learned From Human Gaze
Chen, Y., Liu, C., Shi, B. E., & Liu, M. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2972868 · Cited by 148
“4: the Students003 dataset [30] (average crowd size 26), NYC Grand Central dataset [31] (average crowd size 30), the ZARA2 dataset [30] (average crowd size 7), the Hotel dataset [32] (average crowd size 6), and the the ETH dataset [32] (average crowd size 6).”
Pedestrian Models for Autonomous Driving Part II: High-Level Models of Human Behavior
Camara, F., Bellotto, N., Coşar, S., Weber, F., Nathanael, D., Althoff, M., Wu, J., Ruenz, J., Dietrich, A., Markkula, G. M., et al. (2020) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.3006767 · Cited by 145
“[155] introduced a linear trajectory avoidance (LTA) model which”
BackgroundMX-LSTM: Mixing Tracklets and Vislets to Jointly Forecast Trajectories and Head Poses
Hasan, I., Setti, F., Tsesmelis, T., Bue, A. D., Galasso, F., & Cristani, M. (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.2018.00635 · Cited by 145
“We also consider the Town Center Social Force model (SF) [59] Linear Trajectory Avoidance (LTA) [40] Vanilla LSTM [3] Social LSTM (S-LSTM) [3] MX-LSTM 0 0.5 Motivating the MX-LSTM: a) analysis between the angle discrepancy ω between head pose and movement, the pedestrain smoothed velocity and the…”
BackgroundMethodInfluentialIvanovic, B., Leung, K., Schmerling, E., & Pavone, M. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3043163 · Cited by 154
“The approaches are evaluated on the real-world ETH [39] and UCY [40] pedestrian datasets, a standard benchmark in the field com- prised of challenging multi-human interaction scenarios.”
BackgroundPreserving Structure in Model-Free Tracking
Zhang, L., & van der Maaten, L. J. P. (2013) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2013.221 · Cited by 144
“Whilst significant progress has been made in tracking specific objects (e.g., in tracking faces [1], humans [2], [3], [4], [5], cars [6], and rigid objects [7]), the development of trackers that work well on arbitrary objects remains hard.”
BackgroundMANTRA: Memory Augmented Networks for Multiple Trajectory Prediction
Marchetti, F., Becattini, F., Seidenari, L., & Bimbo, A. D. (2020) · 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/cvpr42600.2020.00717 · Cited by 149
“Several researchers have focused on trajectories of pedestrians [1, 17, 18, 36, 44], either regarded as individuals or crowds, also exploiting social behaviors and interactivity between individuals [1, 17, 18, 29, 36].”
BackgroundContext-Based Path Prediction for Targets with Switching Dynamics
Kooij, J. F. P., Flohr, F. B., Pool, E. A. I., & Gavrila, D. M. (2018) · International Journal of Computer Vision · doi:10.1007/s11263-018-1104-4 · Cited by 143
“A related indicator is the minimum future distance between two agents, which like TTC assumes both travel with fixed velocity (Pellegrini et al. 2009; Cara and de Gelder 2015).”
BackgroundMethodRobot Navigation in Crowded Environments Using Deep Reinforcement Learning
Liu, L., Dugas, D., Cesari, G., et al. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341540 · Cited by 141
“Previous efforts to explicitly model pedestrian behaviours show the difficulty associated with predicting the movements of humans [1], [2], [3].”
BackgroundMoving object detection with laser scanners
Mertz, C., Navarro‐Serment, L. E., MacLachlan, R. A., et al. (2012) · Journal of Field Robotics · doi:10.1002/rob.21430 · Cited by 140
Vehicle detection and tracking in wide field-of-view aerial video
Xiao, J., Cheng, H., Sawhney, H., et al. (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.2010.5540151 · Cited by 138
“Recently tracking people in crowd scene has been gotten more attention by researchers [12, 10 , 2]. Pellegrini et. al. proposed dynamic social behavior model to regulate crowds moving.”
BackgroundTrajectory Unified Transformer for Pedestrian Trajectory Prediction
Shi, L., Wang, L., Zhou, S., et al. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00887 · Cited by 137
“We evaluate TUTR on the most popular datasets for pedestrian trajectory prediction, ETH [19], UCY [14], and SDD [22].”
MethodPedestrian trajectory prediction with convolutional neural networks
Zamboni, S., Kefato, Z. T., Girdzijauskas, Š., et al. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108252 · Cited by 136
“The second metric used is the Final Displacement Error (FDE), which was also introduced in [35].”
MethodInfluentialMultiple Target Tracking in World Coordinate with Single, Minimally Calibrated Camera
Choi, W., & Savarese, S. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15561-1_40 · Cited by 132
“Multi-target tracking can be also aided by considering interaction between targets [13,14,15].”
BackgroundSeeing People in Social Context: Recognizing People and Social Relationships
Wang, G., Gallagher, A., Luo, J., et al. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15555-0_13 · Cited by 132
“In this sense, it is similar to the context of [5, 16].”
BackgroundImproving multi-target tracking via social grouping
Qin, Z., & Shelton, C. R. (2012) · 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.2012.6247899 · Cited by 132
“Sociology and current computer vision research results show that people tend to walk in groups, when they usually stay close to each other and have similar speeds and trajectories [18].”
BackgroundGoal-GAN: Multimodal Trajectory Prediction Based on Goal Position Estimation
Dendorfer, P., Ošep, A., & Leal-Taixé, L. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-69532-3_25 · Cited by 140
“This approach was successfully applied in the domain of multi-object tracking [13,14,15,16].”
MethodOn pairwise costs for network flow multi-object tracking
Chari, V., Lacoste-Julien, S., Laptev, I., & Šivic, J. (2015) · 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.2015.7299193 · Cited by 132
“While linear terms are both simple and easy to minimize, higher order models can represent more useful properties [21].”
BackgroundMethodInfluentialLearning Collective Crowd Behaviors with Dynamic Pedestrian-Agents
Zhou, B., Tang, X., & Wang, X. (2014) · International Journal of Computer Vision · doi:10.1007/s11263-014-0735-3 · Cited by 127
“…Helbing et al, 2000) for crowd simulation, was introduced to the computer vision community recently and was applied to multi-target tracking (Scovanner and Tap-pen, 2009; Pellegrini et al, 2009), abnormality detection (Mehran et al, 2009), and interaction analysis (Scovanner and Tap-pen, 2009).”
BackgroundMethodDetecting F-formations as dominant sets
Hung, H., & Kröse, B. (2011) · INTERNATIONAL CONFERENCE ON MULTIMODAL INTERACTION · doi:10.1145/2070481.2070525 · Cited by 127
A statistical similarity measure for aggregate crowd dynamics
Guy, S. J., van den Berg, J., Liu, W., et al. (2012) · ACM Transactions on Graphics · doi:10.1145/2366145.2366209 · Cited by 126
“The trajectories of each agent were extracted using multi-object tracking [Pellegrini et al. 2009] (see Fig.”
BackgroundMethodHuman identity recognition in aerial images
Oreifej, O., Mehran, R., & Shah, M. A. (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.2010.5540147 · Cited by 124
“In view of the fact that solving recognition (matching) over consecutive frames can be used in tracking, we applied the proposed recognition system on a video and compared the results to the state of the art tracking proposed in [13], which applies a cross correlation matching supported by a social behavior-based mode…”
Group LSTM: Group Trajectory Prediction in Crowded Scenarios
Bisagno, N., Zhang, B., & Conci, N. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-11015-4_18 · Cited by 131
“In the same way as other works [27, 2], we evaluate our results with the following two metrics:”
BackgroundMethodResultGSM: Graph Similarity Model for Multi-Object Tracking
Liu, Q., Chu, Q., Liu, B., & Yu, N. (2020) · Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/74 · Cited by 123
“Many early works extracted hand-craft appearance feature such as raw pixel template [ Yamaguchi et al. , 2011; Pellegrini et al. , 2009 ] , color histogram [ Izadinia et al. , 2012; Xiang et al. , 2015 ] , HOG [ Izadinia et al. , 2012; Kuo et al. , 2010 ] and so on.”
BackgroundTracking Interacting Objects Using Intertwined Flows
Wang, X., Türetken, E., Fleuret, F., & Fua, P. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2513406 · Cited by 120
Knowledge Transfer for Scene-Specific Motion Prediction
Ballan, L., Castaldo, F., Alahi, A., Palmieri, F., & Savarese, S. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46448-0_42 · Cited by 115
“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].”
LAformer: Trajectory Prediction for Autonomous Driving with Lane-Aware Scene Constraints
Liu, M., Cheng, H., Chen, L., Broszio, H., Li, J., Zhao, R., Sester, M., & Yang, M. Y. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00209 · Cited by 117
“Agents are interconnected for social connections and collision avoidance [33, 17].”
BackgroundSocially Constrained Structural Learning for Groups Detection in Crowd
Solera, F., Calderara, S., & Cucchiara, R. (2015) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2015.2470658 · Cited by 114
“We selected two publicly available datasets, namely the BIWI Walking Pedestrians dataset [45] and the Crowds-ByExamples (CBE) dataset [46].”
Occlusion Geodesics for Online Multi-object Tracking
Possegger, H., Mauthner, T., Roth, P. M., & Bischof, H. (2014) · 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.2014.170 · Cited by 115
Wang, B., Wang, L., Shuai, B., Zuo, Z. C., Liu, T., Chan, K. L., & Wang, G. (2016) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2016.55 · Cited by 111
“[31] 65.”
Pfeiffer, M., Paolo, G., Sommer, H., Nieto, J. I., Siegwart, R., & Cadena, C. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8461157 · Cited by 116
“We use a publicly available pedestrian tracking dataset [11] to evaluate our model on real-world data.”
BackgroundMethodInfluentialEvaluation of Socially-Aware Robot Navigation
Gao, Y., & Huang, C. (2022) · Frontiers in Robotics and AI · doi:10.3389/frobt.2021.721317 · Cited by 114
“As opposed to most simulation-based approaches where agent behaviors are generated from crowd simulation [e.g., using Optimal Reciprocal Collision Avoidance (ORCA) (Van Den Berg et al., 2011) or Social Force Model (SFM) (Helbing and Molnár, 1995)], human behaviors in SocNavBench are grounded in real-world datasets (i.…”
MethodInfluentialBRVO: Predicting pedestrian trajectories using velocity-space reasoning
Kim, S., Guy, S. J., Liu, W., et al. (2014) · The International Journal of Robotics Research · doi:10.1177/0278364914555543 · Cited by 114
“After the training frames, we assume that no further sensor information is given and predict the paths using only the motion models.”
BackgroundMethodResultInfluentialA Review of Deep Learning-Based Methods for Pedestrian Trajectory Prediction
Sighencea, B. I., Stanciu, R., & Căleanu, C. D. (2021) · Sensors · doi:10.3390/s21227543 · Cited by 119
“From now onward, these two datasets will be mentioned as the ETH-UCY dataset, considering the fact that they are used jointly.”
BackgroundMethodInfluentialOnline Learned Discriminative Part-Based Appearance Models for Multi-human Tracking
Yang, B., & Nevatia, R. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33718-5_35 · Cited by 112
Tracking Sports Players with Context-Conditioned Motion Models
Liu, J., Carr, P., Collins, R. T., & Liu, Y. (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.2013.239 · Cited by 112
“Recent work in pedestrian tracking has demonstrated promising results by formulating multi-target tracking in terms of data association [1, 4, 7, 18, 23, 25, 27, 29]: a set of potential target locations are estimated in each frame using an object detector, and target trajectories are inferred by…”
BackgroundTrajectory prediction for heterogeneous traffic-agents using knowledge correction data-driven model
Xu, X., Liu, W., & Yu, L. (2022) · Information Sciences · doi:10.1016/j.ins.2022.06.073 · Cited by 112
Yang, D., Li, L., Redmill, K., et al. (2019) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2019.8814092 · Cited by 110
“Existing dataset such as ETH [9] and UCY [10] only covers interpersonal interaction, which is not suitable for VCI.”
MethodInfluentialWe Are Family: Joint Pose Estimation of Multiple Persons
Eichner, M. L., & Ferrari, V. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15549-9_17 · Cited by 109
“In [14] the authors learn the behavior of people in crowded urban scenes and predict the path of a pedestrian given the location and direction of others.”
BackgroundYang, B., Fan, F., Ni, R., et al. (2024) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3345296 · Cited by 109
“Both quantitative and qualitative evaluations of ETH [16], UCY [17], and ApolloScape [18] are conducted.”
Method3DOF Pedestrian Trajectory Prediction Learned from Long-Term Autonomous Mobile Robot Deployment Data
Sun, L., Yan, Z., Mellado, S. M., Hanheide, M., & Duckett, T. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8461228 · Cited by 109
“Following the settings of previous research [2], [3], [7], [8], [9], we sample the frames at 2.”
BackgroundMethodInfluentialA survey on human-aware robot navigation
Möller, R., Furnari, A., Battiato, S., Härmä, A., & Farinella, G. M. (2021) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2021.103837 · Cited by 120
“The dataset contains videos from two di ff erent vantage points and was used for the ICCV 2009 paper [306] that presents a solution to the task of multi-person tracking in crowded environments.”
MethodMotion Influence Map for Unusual Human Activity Detection and Localization in Crowded Scenes
Lee, D.-G., Suk, H., Park, S.-K., & Lee, S. (2015) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2015.2395752 · Cited by 104
“interest points to explain whether they will meet in the near future [28].”
BackgroundMethodEiffert, S., Li, K., Shan, M., et al. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3004324 · Cited by 104
“We deal with the ETH-Univ frame rate issue addressed by [24] similarly by treating every 6 frames as 0.4s rather than 10 frames, and retrain all comparative models for this scene.”
BackgroundInfluentialIntent Prediction of Pedestrians via Motion Trajectories Using Stacked Recurrent Neural Networks
Saleh, K., Hossny, M., & Nahavandi, S. (2018) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2018.2873901 · Cited by 103
“The advantages of this formulation is that a long-term prediction about the pedestrians’ intentions, which is essential for fully and highly automated vehicles, could be effectively inferred [22], [23].”
BackgroundGATraj: A graph- and attention-based multi-agent trajectory prediction model
Cheng, H., Liu, M., Lin, C., Broszio, H., Sester, M., & Yang, M. Y. (2023) · ISPRS Journal of Photogrammetry and Remote Sensing · doi:10.1016/j.isprsjprs.2023.10.001 · Cited by 111
Multi-hypothesis motion planning for visual object tracking
Gong, H., Sim, J., Likhachev, M. A., et al. (2011) · International Conference on Computer Vision · doi:10.1109/iccv.2011.6126296 · Cited by 103
“Second, we create more flexible and realistic hypotheses for possible pedestrian trajectories than others; simpler models of dynamic social behavior [10] are limited in expressive power because they use single hypotheses and short-term predictions.”
BackgroundMethodInfluentialZitouni, M., Bhaskar, H., Dias, J., & Al-Mualla, M. (2016) · Neurocomputing · doi:10.1016/j.neucom.2015.12.070 · Cited by 98
Kinects and human kinetics: A new approach for studying pedestrian behavior
Seer, S., Brändle, N., & Ratti, C. (2014) · Transportation Research Part C Emerging Technologies · doi:10.1016/j.trc.2014.08.012 · Cited by 103
“Vision based tracking has seen considerable progress in recent years, with current systems able to track people through long and challenging sequences (Pellegrini et al., 2009).”
BackgroundMethodInfluentialPositive technology for elderly well-being: A review
Grossi, G., Lanzarotti, R., Napoletano, P., Noceti, N., & Odone, F. (2019) · Pattern Recognition Letters · doi:10.1016/j.patrec.2019.03.016 · Cited by 111
Ghorai, P., Eskandarian, A., Kim, Y., & Mehr, G. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3160932 · Cited by 110
Assessing team strategy using spatiotemporal data
Lucey, P., Oliver, D., Carr, P., et al. (2013) · ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/2487575.2488191 · Cited by 99
Adaptive Trajectory Prediction via Transferable GNN
Xu, Y., Wang, L., Wang, Y., & Fu, Y. (2022) · 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/cvpr52688.2022.00641 · Cited by 102
“Experiments are conducted on two real-world datasets: ETH [56] and UCY [36] as these two public datasets are widely used in this task.”
MethodZen, G., Lepri, B., Ricci, E., & Lanz, O. (2010) · doi:10.1145/1878039.1878048 · Cited by 99
PANDA: A Gigapixel-Level Human-Centric Video Dataset
Wang, X., Zhang, X., Zhu, Y., Guo, Y., Yuan, X., Xiang, L., Wang, Z., Ding, G., Brady, D. J., Dai, Q., et al. (2020) · 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/cvpr42600.2020.00333 · Cited by 107
“On the other hand, to investigate pedestrians in surveillance perspectives, UCY Crowds-by-Example [45], ETH BIWI Walking Pedestrians [55], Town Center [5] and Train Station [81] are proposed for trajectory prediction, abnormal behaviour detection, and pedestrian motion analysis.”
BackgroundMethodInfluentialOnline Multi-Object Tracking Using Joint Domain Information in Traffic Scenarios
Tian, W., Lauer, M., & Chen, L. F. (2019) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2019.2892413 · Cited by 97
“[14] observe”
StarNet: Pedestrian Trajectory Prediction using Deep Neural Network in Star Topology
Zhu, Y., Qian, D., Ren, D., et al. (2019) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros40897.2019.8967811 · Cited by 97
“We evaluate our model on two human crowded trajectory datasets: ETH [24] and UCY [25].”
MethodScenario Understanding and Motion Prediction for Autonomous Vehicles—Review and Comparison
Karle, P., Geisslinger, M., Betz, J., & Lienkamp, M. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3156011 · Cited by 107
“The model is evaluated on multiple real-world datasets for vehicles [87] and pedestrians [135], [156].”
BackgroundMethodMohamed, A., Zhu, D., Vu, W., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20047-2_27 · Cited by 97
Trace and Pace: Controllable Pedestrian Animation via Guided Trajectory Diffusion
Rempe, D., Luo, Z., Peng, X. B., Yuan, Y., Kitani, K., Kreis, K., Fidler, S., & Litany, O. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01322 · Cited by 97
“Since there are pedestrian datasets with diverse motions but no semantic maps [29, 41], and others with limited motions but detailed maps [4], we find mixed training is beneficial to boost diversity and controllability (see Sec.”
BackgroundMethodInfluentialTracking Interacting Objects Optimally Using Integer Programming
Wang, X., Türetken, E., Fleuret, F., & Fua, P. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10590-1_2 · Cited by 95
“In [20,28], group behavior is considered during the tracking process by including priors that account for the fact that people tend to avoid hitting each other and sometimes walk in groups.”
BackgroundEgocentric Future Localization
Park, H. S., Hwang, J.-J., Niu, Y., & Shi, J. (2016) · 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.2016.508 · Cited by 94
Non-Markovian Globally Consistent Multi-object Tracking
Maksai, A., Wang, X., Fleuret, F., & Fua, P. (2017) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.278 · Cited by 94
“However, it has been pointed out that MOTA does not properly account for identity switches [10, 94, 12], as depicted on the left side of Fig.”
BackgroundScene-LSTM: A Model for Human Trajectory PredictionPreprint
Manh, H. T., & Alaghband, G. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1808.04018 · Cited by 137
A survey on human activity recognition from videos
Subetha, T., & Chitrakala, S. (2016) · International Conference on Information Communication and Embedded Systems · doi:10.1109/icices.2016.7518920 · Cited by 95
“The assumption made in social interaction model depicted in [23] and [24] is people react habitually while walking in groups.”
BackgroundModeling cooperative navigation in dense human crowds
Vemula, A., Muelling, K., & Oh, J. (2017) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2017.7989199 · Cited by 94
“We evaluate our model on a publicly available human-trajectory dataset released by ETH, [11].”
BackgroundMethodInfluentialScePT: Scene-consistent, Policy-based Trajectory Predictions for Planning
Chen, Y., Ivanovic, B., & Pavone, M. (2022) · 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/cvpr52688.2022.01659 · Cited by 92
“In particular, we make use of the well-known ETH [35], UCY [30], and nuScenes [7] datasets.”
MethodInfluentialTowards Computational Proxemics: Inferring Social Relations from Interpersonal Distances
Cristani, M., Paggetti, G., Vinciarelli, A., et al. (2011) · doi:10.1109/passat/socialcom.2011.32 · Cited by 91
“The proposed approach includes two main stages: the first is the detection of F-formations, and the second is the inference of social relations from interpersonal distances.”
BackgroundDenseCAvoid: Real-time Navigation in Dense Crowds using Anticipatory Behaviors
Sathyamoorthy, A. J., Liang, J., Patel, U., Guan, T., Chandra, R., & Manocha, D. (2020) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra40945.2020.9197379 · Cited by 89
“Social Force model [34], LTA [35], and ATTR [36] are other non-linear motion models that have been used for pedestrian tracking in low to medium density crowds.”
MethodComputer vision and long short-term memory: Learning to predict unsafe behaviour in construction
Kong, T., Fang, W., Love, P. E. D., Luo, H., Xu, S., & Li, H. (2021) · Advanced Engineering Informatics · doi:10.1016/j.aei.2021.101400 · Cited by 90
Human trajectory prediction and generation using LSTM models and GANs
Rossi, L., Paolanti, M., Pierdicca, R., & Frontoni, E. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108136 · Cited by 93
Learning Fine-Grained Spatial Models for Dynamic Sports Play Prediction
Yue, Y., Lucey, P., Carr, P., et al. (2014) · Proceedings · doi:10.1109/icdm.2014.106 · Cited by 87
“Compared to more typical settings for analyzing spatiatemporal data such crowd and surveillance domains [12], [13], analyzing behaviors in team sports is typically more complex as it is adversarial compared to cooperative, and is contingent on the interaction between and within groups compared to…”
BackgroundScribble Tracker: A Matting-Based Approach for Robust Tracking
Fan, J., Shen, X., & Wu, Y. (2011) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2011.257 · Cited by 88
“In the perspective of model design and update in tracking, early works tended to construct the model by describing the target itself [11], [7], [36], [23], while recently the adoption of context information has become very popular [43], [10], [32], [1], [14], [28], [44], [48], [34], [47].”
BackgroundDecentralized particle filter for joint individual-group tracking
Bazzani, L., Cristani, M., & Murino, V. (2012) · 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.2012.6247888 · Cited by 89
“Thus, intersection operations among bounding boxes translate naturally in that of convex hulls.”
BackgroundMethodInfluentialUnderstanding High-Level Semantics by Modeling Traffic Patterns
Zhang, H., Geiger, A., & Urtasun, R. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.379 · Cited by 87
Zhou, H., Ouyang, W., Cheng, J., Wang, X., & Li, H. (2018) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2018.2825679 · Cited by 84
“For online MOT methods, there were investigations on modeling inter-object interactions with social force models [5], [6], [7], relative spatial and speed differences [8], [9], [10], and relative motion constraints [3], [11].”
MethodDiscriminative Appearance Modeling with Multi-track Pooling for Real-time Multi-object Tracking
Kim, C., Li, F., Alotaibi, M. B., & Rehg, J. M. (2021) · 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/cvpr46437.2021.00943 · Cited by 89
“[28] incorporated a social behavior model into their tracking algorithm.”
BackgroundMethodInteracting Tracklets for Multi-Object Tracking
Lan, L., Wang, X., Zhang, S. S., Tao, D., Gao, W., & Huang, T. S. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2843129 · Cited by 86
“In recent years, there are several approaches [14], [16], [18], [23], [24], [33], [36], [37], [42], [43], [46], [50] that conduct MOT by exploiting the interactions among multiple targets.”
BackgroundProbabilistic group-level motion analysis and scenario recognition
Chang, M., Krahnstoever, N. O., & Ge, W. (2011) · International Conference on Computer Vision · doi:10.1109/iccv.2011.6126312 · Cited by 85
“Existing methods include using social force model [20, 18], floor field [1], correlated topic model [21], and use multi-camera tracking [8, 6].”
MethodRepresenting and Discovering Adversarial Team Behaviors Using Player Roles
Lucey, P., Bialkowski, A., Carr, P., Morgan, S., Matthews, I. A., & Sheikh, Y. A. (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.2013.349 · Cited by 85
“Most of the research in this area has focussed on multi-agent tracking [5, 23] and anomaly detection [32].”
BackgroundInfluentialData-driven crowd evacuation: A reinforcement learning method
Yao, Z., Zhang, G., Lu, D., & Liu, H. (2019) · Neurocomputing · doi:10.1016/j.neucom.2019.08.021 · Cited by 86
Social and Scene-Aware Trajectory Prediction in Crowded Spaces
Lisotto, M., Coscia, P., & Ballan, L. (2019) · doi:10.1109/iccvw.2019.00314 · Cited by 85
“Navigation map allows better predicted trajectories on ETH dataset compared to HOTEL dataset.”
BackgroundMethodInfluentialTrajectron++: Multi-Agent Generative Trajectory Forecasting With Heterogeneous Data for Control
Salzmann, T., Ivanovic, B., Chakravarty, P., & Pavone, M. (2020) · Cited by 85
Karamouzas, I., Sohre, N. E., Narain, R., et al. (2017) · ACM Transactions on Graphics · doi:10.1145/3072959.3073705 · Cited by 85
“Karamouzas et al. [2014] have recently analyzed a wide range of publicly available crowd data [Lerner et al. 2007; Pellegrini et al. 2009; Seyfried et al. 2009], showing that the interaction energy between any given pair of pedestrians follows a power law as a function of their projected time to…”
BackgroundMulti Object Tracking with UAVs using Deep SORT and YOLOv3 RetinaNet Detection Framework
Kapania, S., Saini, D., Goyal, S., Thakur, N., Jain, R., & Nagrath, P. (2020) · doi:10.1145/3377283.3377284 · Cited by 85
“Typical appearance models may employ single [1, 34, 45] or multiple cues.”
BackgroundSocially and Contextually Aware Human Motion and Pose Forecasting
Adeli, V., Adeli, E., Reid, I. D., Niebles, J. C., & Rezatofighi, H. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.3010742 · Cited by 85
“Traditional models use hand-crafted features and rules such as “social forces” in order to model and predict human social interactions [23], [45]–[49].”
MethodThree Steps to Multimodal Trajectory Prediction: Modality Clustering, Classification and Synthesis
Sun, J., Li, Y., Fang, H.-S., & Lu, C. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01300 · Cited by 85
“Performance of our method is evaluated on popular datasets, including ETH [24]/UCY [12] Dataset and Stanford Drone Dataset [26].”
MethodPedestrian Behavior Prediction Using Deep Learning Methods for Urban Scenarios: A Review
Zhang, C., & Berger, C. (2023) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2023.3281393 · Cited by 84
View Vertically: A Hierarchical Network for Trajectory Prediction via Fourier Spectrums
Wong, C., Xia, B., Hong, Z., et al. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20047-2_39 · Cited by 83
“Results in Table 5 favorably demonstrate the performance gains (12.9% on ADE and 11.1% on FDE) by using trajectory spectrums, independent of the rest of the designs in the proposed network.”
InfluentialPedestrian Behavior Modeling From Stationary Crowds With Applications to Intelligent Surveillance
Yi, S., Li, H., & Wang, X. (2016) · IEEE Transactions on Image Processing · doi:10.1109/tip.2016.2590322 · Cited by 71
“In order to better demonstrate the effectiveness of the proposed walking path prediction model, three existing methods are compared, including the Social Force Model (SFM) [18], the Linear Trajectory Avoidance model (LTA) [41], and the Temporal Information Modeling based prediction (TIM) [47].”
MethodTracklet Association by Online Target-Specific Metric Learning and Coherent Dynamics Estimation
Wang, B., Wang, G., Chan, K. L., & Wang, L. (2016) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2551245 · Cited by 81
“- - 231 58.6% 34.4% 7.0% 363 165 GMCP-Tracker [61] 75.6% 71.9% - - - 231 - - - - - MCNF [62] 69.1% 72.0% - - - 231 53.0% 37.7 9.3% 440 243 SMT [8] 64.3% 80.2% - - - 231 67.4% 26.1% 6.5% 343 222 MSBMT [43] 65.5% 71.8% - - - 231 59.1% 33.9% 7.0% 499 288 OMAT [53] 69.5% 68.7% - - - 231 64.7% 27.4% 7.9% 453 209 WAYWAG [55…”
InfluentialRejecting Motion Outliers for Efficient Crowd Anomaly Detection
Khan, M. U. K., Park, H.-S., & Kyung, C. (2018) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2018.2856189 · Cited by 80
“Similarly, the codebook model is used to detect anomalies in [22].”
MethodWhere are you heading? Dynamic Trajectory Prediction with Expert Goal Examples
Zhao, H., & Wildes, R. P. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00753 · Cited by 83
Non-Probability Sampling Network for Stochastic Human Trajectory Prediction
Bae, I., Park, J.-H., & Jeon, H. (2022) · 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/cvpr52688.2022.00637 · Cited by 80
“Many earlier works [15, 38, 43, 58] on human trajectory prediction are based on deterministic approaches which yield the * Corresponding author Figure 1.”
BackgroundMethodInfluentialBayesian intention inference for trajectory prediction with an unknown goal destination
Best, G., & Fitch, R. (2015) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2015.7354203 · Cited by 79
“If the end position is known then this information can be used to improve the predictions [11].”
Background
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
