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.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
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 · see citing works
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 · see citing works
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 · see citing works
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 2,911 · see citing works
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 · see citing works
Lerner, A., Chrysanthou, Y., & Lischinski, D. (2007) · Computer Graphics Forum · doi:10.1111/j.1467-8659.2007.01089.x · Cited by 1,199 · see citing works
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 1,022 · see citing works
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 · see citing works
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 · see citing works
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 · see citing works
Wu, B., & Nevatia, R. (2007) · International Journal of Computer Vision · doi:10.1007/s11263-006-0027-7 · Cited by 764 · see citing works
A mobile vision system for robust multi-person tracking
Ess, A., Leibe, B., Schindler, K., & Van Gool, L. (2008) · 2008 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2008.4587581 · Cited by 721 · see citing works
Self-citationRobust tracking-by-detection using a detector confidence particle filter
Breitenstein, M. D., Reichlin, F., Leibe, B., Koller-Meier, E., & Gool, L. V. (2009) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2009.5459278 · Cited by 555 · see citing works
SPECIFICATION OF THE SOCIAL FORCE PEDESTRIAN MODEL BY EVOLUTIONARY ADJUSTMENT TO VIDEO TRACKING DATA
Johansson, A. F., Helbing, D., & Shukla, P. K. (2007) · Advances in Complex Systems · doi:10.1142/s0219525907001355 · Cited by 513 · see citing works
Robust Object Tracking by Hierarchical Association of Detection Responses
Huang, C., Wu, B., & Nevatia, R. (2008) · Lecture notes in computer science · doi:10.1007/978-3-540-88688-4_58 · Cited by 501 · see citing works
Unsupervised Bayesian Detection of Independent Motion in Crowds
Brostow, G., & Cipolla, R. (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.2006.320 · Cited by 408 · see citing works
Floor Fields for Tracking in High Density Crowd Scenes
Ali, S., & Shah, M. A. (2008) · Lecture notes in computer science · doi:10.1007/978-3-540-88688-4_1 · Cited by 388 · see citing works
Coupled Object Detection and Tracking from Static Cameras and Moving Vehicles
Leibe, B., Schindler, K., Cornelis, N., & Gool, L. V. (2008) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2008.170 · Cited by 354 · see citing works
Self-citationTracking multiple humans in crowded environment
Zhao, T., & Nevatia, R. (2004) · 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.2004.1315192 · Cited by 340 · see citing works
Behavioral Priors for Detection and Tracking of Pedestrians in Video Sequences
Antonini, G., Venegas-Martinez, S., Bierlaire, M., & Thiran, J. (2006) · International Journal of Computer Vision · doi:10.1007/s11263-005-4797-0 · Cited by 122 · see citing works
A Unified Framework for Tracking through Occlusions and across Sensor Gaps
Kaucic, R., Perera, A. G. A., Brooksby, G. W., et al. (2005) · 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.2005.53 · Cited by 104 · see citing works
Large-Scale Agent-Based Pedestrian Simulation
Klügl, F., & Rindsfüser, G. (2007) · Lecture notes in computer science · doi:10.1007/978-3-540-74949-3_13 · Cited by 71 · see citing works
Space Syntax Based Agent Simulation
Penn, A. I., & Turner, A. (2001) · Cited by 56 · see citing works
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
Social LSTM: Human Trajectory Prediction in Crowded Spaces
Alahi, A., Goel, K., Ramanathan, V., Robicquet, A., Fei-Fei, L., & Savarese, S. (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.2016.110 · Cited by 3,584 · see citing works
“Finally, we demonstrate that our Social-LSTM is capable of predicting trajectories of pedestrians much more accurately than state-of-the-art methods on two publicly available datasets: ETH [49], and UCY [39].”
BackgroundMethodResultInfluentialTrajectron++: Dynamically-Feasible Trajectory Forecasting with Heterogeneous Data
Salzmann, T., Ivanovic, B., Chakravarty, P., & Pavone, M. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58523-5_40 · Cited by 1,150 · see citing works
“Our method is evaluated on three publicly-available datasets: The ETH [37], UCY [30], and nuScenes [6] datasets.”
MethodSoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints
Sadeghian, A., Kosaraju, V., Sadeghian, A., Hirose, N., Rezatofighi, H., & Savarese, S. (2019) · 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.2019.00144 · Cited by 1,111 · see citing works
“Traditional models for modeling and predicting human-human interactions used “social forces” to capture human motion patterns [9, 18, 31, 22, 2, 23, 21, 17].”
BackgroundMultiple object tracking: A literature review
Luo, W., Xing, J., Milan, A., Zhang, X.-Q., Liu, W., Zhao, X., & Kim, T.-K. (2014) · Artificial Intelligence · doi:10.1016/j.artint.2020.103448 · Cited by 1,095 · see citing works
“Employed Forces [2] repulsion, constancy, fidelity [80] repulsion, constancy, attraction, coherence [58] coherence”
BackgroundMethodDESIRE: Distant Future Prediction in Dynamic Scenes with Interacting Agents
Lee, N., Choi, W., Vernaza, P., Choy, C., Torr, P. H. S., & Chandraker, M. 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.2017.233 · Cited by 1,091 · see citing works
“Predicting the dynamics of multiple objects is also studied in [24, 25, 3, 31], to name a few.”
BackgroundMohamed, A. A., Qian, K., Elhoseiny, M., & Claudel, C. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.01443 · Cited by 1,100 · see citing works
“The model is trained on two human trajectory prediction datasets: ETH [21] and UCY [11].”
MethodTrackFormer: Multi-Object Tracking with Transformers
Meinhardt, T., Kirillov, A., Leal-Taixe, L., & Feichtenhofer, C. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00864 · Cited by 1,081 · see citing works
“Motion can be modelled for trajectory prediction [1, 25, 42] using a constant velocity assumption (CVA) [2, 9] or the social force model [25, 34, 43, 58].”
MethodRecurrent Network Models for Human Dynamics
Fragkiadaki, K., Levine, S., Felsen, P., & Malik, J. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.494 · Cited by 904 · see citing works
“Similarly, for Computer Vision algorithms, predicting human motion is important for timely human-computer interaction [17], obstacle avoidance [22], and people tracking [8].”
BackgroundAutonomous Vehicles That Interact With Pedestrians: A Survey of Theory and Practice
Rasouli, A., & Tsotsos, J. K. (2019) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2019.2901817 · Cited by 884 · see citing works
“treated as a dynamic navigation problem in which pedestrians choose the path that minimizes the likelihood of colliding with others [143].”
BackgroundHuman motion trajectory prediction: a survey
Rudenko, A., Palmieri, L., Herman, M., Kitani, K. M., Gavrila, D. M., & Arras, K. O. (2020) · The International Journal of Robotics Research · doi:10.1177/0278364920917446 · Cited by 789 · see citing works
“Motion state (position and possibly velocity), e.g. (Ferrer and Sanfeliu 2014; Elfring et al. 2014; Pellegrini et al. 2009; Kitani et al. 2012; Karasev et al. 2016; Ziebart et al. 2009; Kooij et al. 2018; Trautman and Krause 2010; Kuderer et al. 2012; Bennewitz et al. 2005; Kucner et al. 2017; Bera…”
BackgroundMethodInfluentialMOTChallenge 2015: Towards a Benchmark for Multi-Target TrackingPreprint
Leal-Taixé, L., Milan, A., Reid, I. D., Roth, S., & Schindler, K. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1504.01942 · Cited by 923 · see citing works
Self-citationUA-DETRAC: A New Benchmark and Protocol for Multi-Object Detection and Tracking
Wen, L., Du, D., Cai, Z., Lei, Z., Chang, M.-C., Qi, H., Lim, J., Yang, M.-H., & Lyu, S. (2020) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2020.102907 · Cited by 804 · see citing works
“Early multi-object tracking methods formulate the task as the state estimation problem using Kalman (Leven and Lanter-man, 2009; Pellegrini et al., 2009) and particle filters (Isard and Blake, 1998; Khan et al., 2005; Mikami et al., 2009; Yang et al., 2014).”
MethodSocial Attention: Modeling Attention in Human Crowds
Vemula, A., Muelling, K., & Oh, J. (2018) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2018.8460504 · Cited by 709 · see citing works
“1) Average Displacement Error: Similar to the metric used in [23], this computes the mean euclidean distance over all estimated points at each time-step in the”
BackgroundMethodInfluentialSTGAT: Modeling Spatial-Temporal Interactions for Human Trajectory Prediction
Huang, Y., Bi, H., Li, Z., Mao, T., & Wang, Z. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00637 · Cited by 697 · see citing works
AgentFormer: Agent-Aware Transformers for Socio-Temporal Multi-Agent Forecasting
Yuan, Y., Weng, X., Ou, Y., & Kitani, K. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00967 · Cited by 671 · see citing works
“We evaluate our method on well-established pedestrian datasets, ETH [38] and UCY [28], and an autonomous driving dataset, nuScenes [3].”
MethodComputer Vision for Autonomous Vehicles
Janai, J., Güney, F., Behl, A., & Geiger, A. (2020) · Foundations and Trends® in Computer Graphics and Vision · doi:10.1561/0600000079 · Cited by 664 · see citing works
Learning Social Etiquette: Human Trajectory Understanding In Crowded Scenes
Robicquet, A., Sadeghian, A., Alahi, A., & Savarese, S. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46484-8_33 · Cited by 927 · see citing works
“Similar to [18] we model such obstacles as agents with zero velocity.”
BackgroundMethodInfluentialUnfreezing the robot: Navigation in dense, interacting crowds
Trautman, P., & Krause, A. (2010) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2010.5654369 · Cited by 624 · see citing works
“We evaluate our approach on a data set of over 8 minutes of video recorded from above a doorway of a university building at ETH (see [18] for more details of the video collect and how to access the data).”
BackgroundMethodResultOnline Multiperson Tracking-by-Detection from a Single, Uncalibrated Camera
Breitenstein, M. D., Reichlin, F., Leibe, B., Koller-Meier, E., & Gool, L. V. (2010) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2010.232 · Cited by 579 · see citing works
“Additionally, to handle occlusions more robustly, other researchers use 3D information [11], [15], train detectors for individual body parts [45], or apply application-specific motion models [35].”
BackgroundSR-LSTM: State Refinement for LSTM Towards Pedestrian Trajectory Prediction
Zhang, P., Ouyang, W., Zhang, P., Xue, J., & Zheng, N. (2019) · 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.2019.01236 · Cited by 579 · see citing works
“There are two types of metrics for evaluating the performance of trajectory prediction, including the Mean Average Displacement (MAD) error and Final Average Displacement (FAD) error [31] in meters.”
MethodLarge Scale Interactive Motion Forecasting for Autonomous Driving : The Waymo Open Motion Dataset
Ettinger, S., Cheng, S., Caine, B., Liu, C., Zhao, H., Pradhan, S., Chai, Y., Sapp, B., Qi, C., Zhou, Y., et al. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00957 · Cited by 584 · see citing works
“The most common is Average Displacement Error (ADE) [1, 24].”
BackgroundTracking the Untrackable: Learning to Track Multiple Cues with Long-Term Dependencies
Sadeghian, A., Alahi, A., & Savarese, S. (2017) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.41 · Cited by 567 · see citing works
“Many models are based on raw pixel template representation for simplicity [71, 4, 68, 51, 50], while color histogram is the most popular representation for appearance modeling in MOT approaches [9, 35].”
BackgroundMethodInfluentialTransformer Networks for Trajectory Forecasting
Giuliari, F., Hasan, I., Cristani, M., & Galasso, F. (2021) · 2020 25th International Conference on Pattern Recognition (ICPR) · doi:10.1109/icpr48806.2021.9412190 · Cited by 572 · see citing works
“ETH [2] and UCY [3], where performance measures are close to saturation, since leading techniques only report average forecasting errors of ∼20cm across 200m-long pavements.”
BackgroundMethodInfluentialLuo, W., Yang, B., & Urtasun, R. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00376 · Cited by 644 · see citing works
“Some work has also focussed on short term prediction of dynamic objects [7, 22].”
BackgroundLearning a Neural Solver for Multiple Object Tracking
Braso, G., & Leal-Taixe, L. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00628 · Cited by 504 · see citing works
“Data association can be done on a frame-by-frame basis for online applications [10, 20, 51] or track-by-track [7].”
BackgroundWho are you with and where are you going?
Yamaguchi, K., Berg, A. C., Ortiz, L. E., & Berg, T. L. (2011) · doi:10.1109/cvpr.2011.5995468 · Cited by 502 · see citing works
“We use the model described in [8] to capture this effect: Note that this term requires three parameters σ d , σ w , and β .”
BackgroundMethodInfluentialThe inD Dataset: A Drone Dataset of Naturalistic Road User Trajectories at German Intersections
Bock, J., Krajewski, R. W., Moers, T., et al. (2020) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv47402.2020.9304839 · Cited by 497 · see citing works
“In Table I, we compare the inD dataset with current public datasets, which are the most similar to the inD dataset: Crowds UCY/Zara [7], Ko-PER [16], DUT [18], CITR [18], BIWI Hotel/ETH [9], VRU Trajectory Dataset [4] and Stanford Drone [12].”
BackgroundMethodMulti-Agent Tensor Fusion for Contextual Trajectory Prediction
Zhao, T., Xu, Y., Monfort, M., Choi, W., Baker, C. L., Zhao, Y., Wang, Y., & Wu, Y. N. (2019) · 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.2019.01240 · Cited by 492 · see citing works
Social-BiGAT: Multimodal Trajectory Forecasting using Bicycle-GAN and Graph Attention NetworksPreprint
Kosaraju, V., Sadeghian, A., Martín-Martín, R., Reid, I. D., Rezatofighi, S. H., & Savarese, S. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1907.03395 · Cited by 482 · see citing works
MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior PredictionPreprint
Chai, Y., Sapp, B., Bansal, M., & Anguelov, D. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1910.05449 · Cited by 482 · see citing works
“Our model contrasts with popular past approaches which either provide only a single MAP trajectory [1, 2, 9, 10, 11] or an unweighted set of samples via a generative model [3, 4, 6, 12, 13, 14, 15].”
MethodThe Trajectron: Probabilistic Multi-Agent Trajectory Modeling With Dynamic Spatiotemporal Graphs
Ivanovic, B., & Pavone, M. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00246 · Cited by 479 · see citing works
“Both of our methods significantly outperform SGAN on the ETH datasets, the UCY Univ dataset, and on average (P <.001; two-tailed t-test on the difference between our and SGAN’s mean NLL).”
BackgroundMethodInfluentialLearning by Tracking: Siamese CNN for Robust Target Association
Leal-Taixe, L., Canton-Ferrer, C., & Schindler, K. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2016.59 · Cited by 444 · see citing works
“A different line of attack is to develop sophisticated motion models in order to better predict a tracked person’s location, most notably models that include interactions between nearby people [1, 10, 35, 44, 47, 56].”
BackgroundMethodSelf-citationPIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction
Rasouli, A., Kotseruba, I., Kunić, T., & Tsotsos, J. K. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00636 · Cited by 468 · see citing works
“Trajectory-based algorithms rely on observation of pedestrians’ past motion history and predict the location of the pedestrians in the future using contextual information such as 3D depth [14, 34, 39], social interactions [25, 42, 1, 31, 41], the ego-vehicle dynamics and the scene structure [16, 17, 5].”
BackgroundSS-LSTM: A Hierarchical LSTM Model for Pedestrian Trajectory Prediction
Xue, H., Huynh, D. Q., & Reynolds, M. A. (2018) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2018.00135 · Cited by 458 · see citing works
“ETH and UCY: ETH [19] and UCY [15] are two publicly available datasets of pedestrian trajectories that cover challenging movement patterns like walking together and collision avoidance.”
BackgroundMethodInfluentialSocially compliant mobile robot navigation via inverse reinforcement learning
Kretzschmar, H., Spies, M., Sprunk, C., & Burgard, W. (2016) · The International Journal of Robotics Research · doi:10.1177/0278364915619772 · Cited by 454 · see citing works
“The second dataset (Pellegrini et al., 2009) comprises 12 minutes of trajectories of pedestrians interacting in a hotel entrance, leading to 71 composite trajectories with three to five pedestrians each.”
BackgroundTrafficPredict: Trajectory Prediction for Heterogeneous Traffic-Agents
Ma, Y., Zhu, X., Zhang, S., Yang, R., Wang, W., & Manocha, D. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33016120 · Cited by 458 · see citing works
“There are some pedestrian trajectory datasets like ETH (Pellegrini et al. 2009), UCY (Lerner et al. 2007), etc., but such datasets only focus on human crowds without any vehicles.”
BackgroundMethodAre They Going to Cross? A Benchmark Dataset and Baseline for Pedestrian Crosswalk Behavior
Rasouli, A., Kotseruba, I., & Tsotsos, J. K. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.33 · Cited by 436 · see citing works
“The common approach to remedy the problem of predicting road users behavior is to employ dynamic factors, such as trajectory [3] or velocity [30], or the expected final goal of pedestrians [18, 31].”
BackgroundA survey of approaches and trends in person re-identification
Bedagkar-Gala, A., & Shah, S. K. (2014) · Image and Vision Computing · doi:10.1016/j.imavis.2014.02.001 · Cited by 420 · see citing works
“Multiple Object Tracking (MOT) within a single camera's FOV has also been widely researched and many algorithms have been proposed [8] [9] [10] [11] [12] over the past two decades, but sustained tracking under varying observation environments remains an open problem.”
Near-Online Multi-target Tracking with Aggregated Local Flow Descriptor
Choi, W. (2015) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.347 · Cited by 409 · see citing works
“(e.g. A1 = {d1, d10, d23}) during the time of presence; i.e. (V T1 ,DT1 )→ A.”
Improving Data Association by Joint Modeling of Pedestrian Trajectories and Groupings
Pellegrini, S., Ess, A., & Gool, L. V. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15549-9_33 · Cited by 387 · see citing works
“Only some researchers use social structures to improve tracking, most notably in crowded scenarios [7], or by modeling of collision avoidance-behavior [8, 11].”
BackgroundSelf-citationUniversal Power Law Governing Pedestrian Interactions
Karamouzas, I., Skinner, B. J., & Guy, S. J. (2014) · Physical Review Letters · doi:10.1103/physrevlett.113.238701 · Cited by 380 · see citing works
SGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction
Shi, L., Wang, L., Long, C., Zhou, S., Zhou, M., Niu, Z., & Hua, G. (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/cvpr46437.2021.00888 · Cited by 378 · see citing works
“Extensive experimetal results on the ETH [34] and UCY [24] datasets show that our method outperforms all the comparison state-of-the-art works.”
MethodA Unified Framework for Multi-target Tracking and Collective Activity Recognition
Choi, W., & Savarese, S. (2012) · Lecture notes in computer science · doi:10.1007/978-3-642-33765-9_16 · Cited by 359 · see citing works
“Much of the existing literature on activity recognition and tracking [3–11] avoids the complexity of this context-rich approach by seeking to solve the problems in isolation.”
BackgroundVision-Based Analysis of Small Groups in Pedestrian Crowds
Ge, W., Collins, R. T., & Ruback, R. B. (2011) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2011.176 · Cited by 354 · see citing works
Kitani, K. M., Ziebart, B. D., Bagnell, J. A., & Hebert, M. (2012) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-33765-9_15 · Cited by 701 · see citing works
“Previous work [11], [12], has shown that modeling the impact of the social environment, like actions of nearby pedestrians, can improve priors over pedestrian trajectories.”
Panoptic Studio: A Massively Multiview System for Social Interaction Capture
Joo, H., Simon, T., Li, X., Liu, H., Tan, L., Gui, L., Banerjee, S. K., Godisart, T., Nabbe, B. C., Matthews, I. A., et al. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2017.2782743 · Cited by 354 · see citing works
“In other methods, spatial characteristics of human social behaviors are modeled as a prior for multi-target motion tracking [14], [15], [16].”
MethodPRECOG: PREdiction Conditioned on Goals in Visual Multi-Agent Settings
Rhinehart, N., McAllister, R., Kitani, K., & Levine, S. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00291 · Cited by 351 · see citing works
“In contrast to the commonly-used average displacement error (ADE) and final displacement error (FDE) metrics that computes the mean Euclidean error from a batch of samples to a single ground-truth sample [1, 6, 10, 14, 28], minMSD has the desirable property of not penalizing plausible samples that correspond to decisi…”
Density-aware person detection and tracking in crowds
Rodriguez, M. D., Laptev, I., Sivic, J., & Audibert, J.-Y. (2011) · Vision · doi:10.1109/iccv.2011.6126526 · Cited by 382 · see citing works
“In the area of tracking, [4, 16, 22, 23] demonstrate how the tracking accuracy can be improved by the joint modeling of motion of multiple people in the scene.”
BackgroundSocial Ways: Learning Multi-Modal Distributions of Pedestrian Trajectories With GANs
Amirian, J., Hayet, J.-B., & Pettré, J. (2019) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw.2019.00359 · Cited by 341 · see citing works
“As it can be seen, the use of our approach leads to significantly lower prediction errors for the ETH and Hotel experiments, but not on the ZARA experiments.”
MethodInfluentialPeople tracking with human motion predictions from social forces
Luber, M., Stork, J. A., Tipaldi, G. D., et al. (2010) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/robot.2010.5509779 · Cited by 330 · see citing works
“Unlike [14] we consider additional forces from the environment by maintaining a short-term environment model and computing different repulsion forces and physical constraints from static obstacles.”
BackgroundUnderstanding collective crowd behaviors: Learning a Mixture model of Dynamic pedestrian-Agents
Zhou, B., Wang, X., & Tang, X. (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.2012.6248013 · Cited by 327 · see citing works
“[5, 4] for crowd simulation, was introduced to the computer vision community recently and was applied to multi-target tracking [15], abnormality detection [12], and interaction analysis [16].”
BackgroundMethodThe Way They Move: Tracking Multiple Targets with Similar Appearance
Dicle, C., Camps, O., & Sznaier, M. (2013) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2013.286 · Cited by 312 · see citing works
“Often, multi-target tracking is formulated as a network flow problem [35] or variations of it [6, 8, 26, 27, 31].”
BackgroundCore Challenges of Social Robot Navigation: A Survey
Mavrogiannis, C., Baldini, F., Wang, A., Zhao, D., Trautman, P., Steinfeld, A. M., & Oh, J. (2023) · ACM Transactions on Human-Robot Interaction · doi:10.1145/3583741 · Cited by 315 · see citing works
“ETH and Hotel have more linear trajectories and lower pedestrian density when compared to others.”
BackgroundMethodInfluentialDeep Learning Methods for Vessel Trajectory Prediction Based on Recurrent Neural Networks
Capobianco, S., Millefiori, L. M., Forti, N., Braca, P., & Willett, P. (2021) · IEEE Transactions on Aerospace and Electronic Systems · doi:10.1109/taes.2021.3096873 · Cited by 311 · see citing works
“In vehicle trajectory prediction, the model performance is usually assessed with the average displacement error (ADE) [55] on all the predicted positions with respect to”
MethodTraPHic: Trajectory Prediction in Dense and Heterogeneous Traffic Using Weighted Interactions
Chandra, R., Bhattacharya, U., Bera, A., & Manocha, D. (2019) · 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.2019.00868 · Cited by 302 · see citing works
“In these cases, the interaction between agents can be modeled using well-known models based on social forces [19], velocity obstacles [35], or LTA [31].”
BackgroundInfluentialContext-Based Pedestrian Path Prediction
Kooij, J. F. P., Schneider, N., Flohr, F. B., & Gavrila, D. M. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10599-4_40 · Cited by 301 · see citing works
“The related Linear Trajectory Avoidance model [28] for short-term path prediction uses the expected point of closest approach to foreshadow and avoid possible collisions.”
BackgroundInfluentialA Hierarchical Representation for Future Action Prediction
Lan, T., Chen, T.-C., & Savarese, S. (2014) · Lecture notes in computer science · doi:10.1007/978-3-319-10578-9_45 · Cited by 298 · see citing works
“Most of the works are focused on predicting (or forecasting) the future trajectories of pedestrians [15,7].”
BackgroundEncoding Crowd Interaction with Deep Neural Network for Pedestrian Trajectory Prediction
Xu, Y., Piao, Z., & Gao, S. (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.2018.00553 · Cited by 286 · see citing works
“As shown in [18], these datasets also cover very challenging group behaviors such as couples walking together, groups crossing each other and groups forming and dispersing in some scenes.”
BackgroundMethodInfluentialFrom Goals, Waypoints & Paths To Long Term Human Trajectory Forecasting
Mangalam, K., An, Y., Girase, H., & Malik, J. (2021) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.01495 · Cited by 284 · see citing works
“We use a total of three datasets to study Y-net’s performance – the Stanford Drone Dataset (SDD) [33], the Intersection Drone Dataset (InD) [5], and the ETH [31] / UCY [23] forecasting benchmark.”
MethodLearning an Image-Based Motion Context for Multiple People Tracking
Leal-Taixé, L., Fenzi, M., Kuznetsova, A., Rosenhahn, B., & Savarese, S. (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.2014.453 · Cited by 281 · see citing works
“Since odometry information is found automatically and is prone to errors, this severely affects the efficacy of [19] and [25].”
BackgroundMethodInfluentialWhat the Constant Velocity Model Can Teach Us About Pedestrian Motion Prediction
Schöller, C., Aravantinos, V., Lay, F. S., & Knoll, A. (2020) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2020.2969925 · Cited by 281 · see citing works
“prediction, the ETH dataset [6] and the UCY dataset [31].”
MethodHuman Trajectory Forecasting in Crowds: A Deep Learning Perspective
Kothari, P., Kreiss, S., & Alahi, A. (2021) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2021.3069362 · Cited by 271 · see citing works
“In this work, we propose to utilize Layer-wise Relevance Propagation (LRP) [16] to explain the decisions of our trajectory forecasting models.”
MethodTrautman, P., Ma, J., Murray, R. M., & Krause, A. (2015) · The International Journal of Robotics Research · doi:10.1177/0278364914557874 · Cited by 270 · see citing works
“In addition, this joint collision avoidance criteria has been used to improve the data association and target tracking of individuals in human crowds (Pellegrini et al., 2009, 2010; Luber et al., 2010).”
BackgroundMethodResultInfluentialScene-Independent Group Profiling in Crowd
Shao, J., Loy, C. C., & Wang, X. (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.2014.285 · Cited by 262 · see citing works
“Most existing imagery-based crowd analysis methods tend to treat a crowd either as a collection of individuals [10, 23, 27] or as an aggregated whole [3, 8, 24, 17].”
BackgroundPeople reidentification in surveillance and forensics
Vezzani, R., Baltieri, D., & Cucchiara, R. (2013) · ACM Computing Surveys · doi:10.1145/2543581.2543596 · Cited by 254 · see citing works
Learning context for collective activity recognition
Choi, W., Shahid, K., & Savarese, S. (2011) · doi:10.1109/cvpr.2011.5995707 · Cited by 247 · see citing works
“As a critical ingredient for successful activity classification, researches have developed methods for identifying [13] and tracking [26, 27, 11] humans under different conditions as well as accurately estimating their pose [14, 5].”
BackgroundUnderstanding pedestrian behaviors from stationary crowd groups
Yi, S., Li, H., & Wang, X. (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.2015.7298971 · Cited by 246 · see citing works
“A typical example is the social force model [12], which was originally proposed for crowd simulation [11], and then was used in tracking [29], interaction analysis [31], and abnormal event detection [25].”
MethodIdentifying Behaviors in Crowd Scenes Using Stability Analysis for Dynamical Systems
Solmaz, B., Moore, B. E., & Shah, M. A. (2012) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2012.123 · Cited by 245 · see citing works
Abnormal detection using interaction energy potentials
Cui, X., Liu, Q., Gao, M., & Metaxas, D. (2011) · doi:10.1109/cvpr.2011.5995558 · Cited by 242 · see citing works
“The calculation of the interaction energy potentials is inspired by the idea of social behaviors[18]: assuming that people are aware of the positions and velocities of other people at time t .”
BackgroundMethodInfluentialAgreeing to cross: How drivers and pedestrians communicate
Rasouli, A., Kotseruba, I., & Tsotsos, J. K. (2017) · IEEE Intelligent Vehicles Symposium · doi:10.1109/ivs.2017.7995730 · Cited by 242 · see citing works
“These works mainly look at the dynamic factors in the scene such as pedestrians’ trajectories [22] and velocities [23] or try to predict the changes in the behavior of pedestrians crossing as a group [24].”
BackgroundMulti-target tracking by online learning of non-linear motion patterns and robust appearance models
Yang, B., & Nevatia, 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.2012.6247892 · Cited by 230 · see citing works
TPNet: Trajectory Proposal Network for Motion Prediction
Fang, L., Jiang, Q., Shi, J., & Zhou, B. (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/cvpr42600.2020.00683 · Cited by 230 · see citing works
“3) We achieve the state-of-the-art re-sults on the recent large-scale trajectory prediction datasets ETH [35], UCY [27], ApolloScape [32] and Argoverse [8].”
BackgroundMethodRecursive Social Behavior Graph for Trajectory Prediction
Sun, J.-H., Jiang, Q.-H., & Lu, C.-W. (2020) · Computer Vision and Pattern Recognition · doi:10.1109/cvpr42600.2020.00074 · Cited by 191 · see citing works
“Experiments on multiple human trajectory benchmark, including two datasets in ETH[31] and three datasets in UCY[21], show the superior of our model in accuracy improvement.”
BackgroundMethodInfluentialLong-Term Human Motion Prediction with Scene Context
Cao, Z., Gao, H., Mangalam, K., Cai, Q.-Z., Vo, M., & Malik, J. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58452-8_23 · Cited by 227 · see citing works
“nteractions using physical forces [18], continuum dynamics [52], Hidden Markov model [25], or game theory [34]. Many of these approaches achieve competitive results even on modern pedestrian datasets [30,46]. With the resurgence of neural nets, data-driven prediction paradigm that captures multi-modal interaction betw…”
BackgroundSocially-Aware Large-Scale Crowd Forecasting
Alahi, A., Ramanathan, V., & Fei-Fei, L. (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.2014.283 · Cited by 218 · see citing works
“We are interested in behavioral cues that remain stable over time and across various sensing modalities (e.g. optical, thermal, and depth) to link far-away tracklets.”
BackgroundZhou, H., Ren, D., Xia, H., Fan, M., Yang, X., & Huang, H. B. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.03.024 · Cited by 220 · see citing works
Looking at Humans in the Age of Self-Driving and Highly Automated Vehicles
Ohn-Bar, E., & Trivedi, M. M. (2016) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2016.2571067 · Cited by 212 · see citing works
“[113], 2009 I Pedestrian Path Prediction PD, Social Context Kooij et al.”
BackgroundDeep Network Flow for Multi-object Tracking
Schulter, S., Vernaza, P., Choi, W., & Chandraker, M. 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.2017.292 · Cited by 213 · see citing works
Social interaction discovery by statistical analysis of F-formations
Cristani, M., Bazzani, L., Paggetti, G., et al. (2011) · British Machine Vision Conference (BMVC) · doi:10.5244/c.25.23 · Cited by 211 · see citing works
“The social force model has been modified in [27], where SFM is embedded as model for the dynamics in a tracking framework.”
BackgroundMethodLeal-Taixé, L., Pons‐Moll, G., & Rosenhahn, B. (2011) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2011.6130233 · Cited by 212 · see citing works
“In [22] and [17], the social forces are included in the motion model of the Kalman or Extended Kalman filter.”
BackgroundMethodStepwise Goal-Driven Networks for Trajectory Prediction
Wang, C., Wang, Y., Xu, M., & Crandall, D. (2022) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2022.3145090 · Cited by 212 · see citing works
“ETH [44] and UCY [45] include 1,536 pedestrians in 5 sets of data with 4 unique scenes.”
BackgroundInfluentialVideo-Based Human Behavior Understanding: A Survey
Borges, P. V. K., Conci, N., & Cavallaro, A. (2013) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2013.2270402 · Cited by 208 · see citing works
“[158] applied this model to crowded scenes, considering single moving entities as agents.”
MethodGroupNet: Multiscale Hypergraph Neural Networks for Trajectory Prediction with Relational Reasoning
Xu, C., Li, M., Ni, Z., et al. (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/cvpr52688.2022.00639 · Cited by 206 · see citing works
“We evaluate our method on both physical simulation datasets and three real-world public datasets: the NBA SportVU Dataset (NBA), the ETH-UCY Dataset [21, 31], and the Stanford Drone Dataset (SDD) [32].”
MethodBiTraP: Bi-Directional Pedestrian Trajectory Prediction With Multi-Modal Goal Estimation
Yao, Y., Atkins, E. M., Johnson‐Roberson, M., et al. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3056339 · Cited by 205 · see citing works
“Two FPV datasets, Joint Attention for Autonomous Driving (JAAD) [35] and Pedestrian Intention Estimation (PIE) [22], and two benchmark BEV datasets, ETH [36] and UCY [37], were used in our experiments.”
MethodA Streakline Representation of Flow in Crowded Scenes
Mehran, R., Moore, B. E., & Shah, M. A. (2010) · Lecture notes in computer science · doi:10.1007/978-3-642-15558-1_32 · Cited by 200 · see citing works
“[8] in tracking based on social force model.”
Human behavior analysis in video surveillance: A Social Signal Processing perspective
Cristani, M., Ramachandra, R., Bue, A. D., & Murino, V. (2012) · Neurocomputing · doi:10.1016/j.neucom.2011.12.038 · Cited by 198 · see citing works
“Therefore, the design of a socially driven dynamic model for tracking may be the key ingredient to overcome the current limitations of the current algorithms, as already shown in some recent approaches exploiting the Social Force Model [19] [20] (see later for further details) .”
MethodA Literature Review on the Prediction of Pedestrian Behavior in Urban Scenarios
Ridel, D. A., Rehder, E., Lauer, M., Stiller, C., & Wolf, D. F. (2018) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2018.8569415 · Cited by 192 · see citing works
“lowed within a shared common space [40], [41], [42], [43].”
BackgroundStructure Preserving Object Tracking
Zhang, L., & van der Maaten, L. (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.2013.240 · Cited by 189 · see citing works
Context-Aware Trajectory Prediction
Bartoli, F., Lisanti, G., Ballan, L., & Bimbo, A. D. (2018) · Proceedings - International Conference on Pattern Recognition/Proceedings/International Conference on Pattern Recognition · doi:10.1109/icpr.2018.8545447 · Cited by 191 · see citing works
“In fact, being able to take into account the interactions of nearby agents, is extremely important to avoid collisions in a crowded scenario [1], [19], [10].”
BackgroundInfluentialA General Framework for Tracking Multiple People from a Moving Camera
Choi, W., Pantofaru, C., & Savarese, S. (2013) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2012.248 · Cited by 207 · see citing works
“However, recent studies [36], [31], [24], including our own [9], suggest that modeling the interaction between targets enhances tracking accuracy significantly.”
Online Multi-object Tracking via Structural Constraint Event Aggregation
Yoon, J. H., Lee, C., Yang, M., & Yoon, K. (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.2016.155 · Cited by 184 · see citing works
“Pellegrini et al. [21] and Leal-Taixé et al. [18] use social force models which consider pairwise motion (such as attraction and repulsion) and visual odometry to obtain 3D motion information for tracking multiple objects.”
BackgroundKorbmacher, R., & Tordeux, A. (2022) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2022.3205676 · Cited by 183 · see citing works
“In Table VII, the ADE and FDE metrics from selected studies using the social-LSTM as a benchmark are shown.”
BackgroundInfluentialParameter estimation and comparative evaluation of crowd simulations
Wolinski, D., Guy, S. J., Olivier, A., Lin, M. C., Manocha, D., & Pettré, J. (2014) · Computer Graphics Forum · doi:10.1111/cgf.12328 · Cited by 171 · see citing works
Pedestrian Behavior Understanding and Prediction with Deep Neural Networks
Yi, S., Li, H., & Wang, X. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46448-0_16 · Cited by 168 · see citing works
“In order to validate prediction robustness, the proposed method is also evaluated on five more datasets, i.e. ETH [17], Hotel [17], ZARA01 [42], ZARA02 [42], and UCY [42].”
BackgroundMethodInfluentialSocial Interactions for Autonomous Driving: A Review and Perspectives
Wang, W., Wang, L., Zhang, C., Liu, C., & Sun, L. (2022) · Foundations and Trends in Robotics · doi:10.1561/2300000078 · Cited by 168 · see citing works
Joint inference of groups, events and human roles in aerial videos
Shu, T., Xie, D., Rothrock, B., et al. (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.2015.7299088 · Cited by 164 · see citing works
“Unfortunately, popular appearance-based approaches to detecting people and objects used to produce input for recognizing group events and interactions [25, 7, 32, 16, 30, 9] do not handle the above three challenges.”
BackgroundUnderstanding Collective Activitiesof People from Videos
Choi, W., & Savarese, S. (2014) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2013.220 · Cited by 161 · see citing works
“In many application scenarios (such as surveillance, video analysis and retrieval), however, it is critical to analyze the collective behavior of individuals (which we refer as to collective activities) as well as the way such individuals interact [7], [21], [35].”
BackgroundSocialVAE: Human Trajectory Prediction Using Timewise Latents
Xu, P., Hayet, J.-B., & Karamouzas, I. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19772-7_30 · Cited by 159 · see citing works
“Our approach shows state-of-the-art performance in most of the test cases from the ETH/UCY and SDD trajectory prediction benchmarks.”
MethodInfluentialRobot navigation in dense human crowds: the case for cooperation
Trautman, P., Ma, J., Murray, R. M., & Krause, A. (2013) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra.2013.6630866 · Cited by 158 · see citing works
“This prediction model is similar to the state of the art crowd prediction models of [31] and [32].”
ResultPedestrian Trajectory Prediction Based on Deep Convolutional LSTM Network
Song, X., Chen, K., Li, X., et al. (2020) · IEEE Transactions on Intelligent Transportation Systems · doi:10.1109/tits.2020.2981118 · Cited by 158 · see citing works
“These datasets To compute prediction errors, we consider the following two metrics: 1)Average Displacement Error (ADE): Similar to the metric used in [56], this computes the mean Euclidean distance over all estimated points at each time-step in the predicted trajectory and true trajectory.”
BackgroundMethod
- 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
- Physical Sciences › Engineering › Ocean Engineering › 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
