Pith. sign in

REVIEW 1 major objections 1 minor 47 references

A new annotation framework extracts trajectory data for dense pedestrian-vehicle interactions from uncalibrated surveillance videos in unstructured scenes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.3

2026-06-29 22:24 UTC pith:G67SM7CI

load-bearing objection PINNS dataset release targets unstructured scene data gap but lacks reported validation on trajectory annotations from uncalibrated cameras. the 1 major comments →

arxiv 2605.25947 v1 pith:G67SM7CI submitted 2026-05-25 cs.CV

A Pedestrian-Vehicle Interaction Benchmark and Annotation Framework for Unstructured Scenes via Uncalibrated Cameras

Haoyang Peng , Qian Hu , Songan Zhang , Ming Yang This is my paper
classification cs.CV
keywords pedestrian-vehicle interactionunstructured scenesuncalibrated camerastrajectory predictionautonomous drivingmixed trafficdatasetannotation framework
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Existing datasets for pedestrian-vehicle interaction prediction come mostly from structured roads and leave complex unstructured environments poorly represented. The paper presents an annotation framework that works directly with video from uncalibrated cameras to build the PINNS dataset. PINNS supplies both trajectory data and scene-level information across multiple countries, seasons, lighting conditions, and weather, with an emphasis on dense mixed-traffic scenes. The authors also review open challenges in heterogeneous agent prediction and position the dataset as a resource to address them. If the framework works as described, researchers gain a public, extensible source of real-world interaction data that structured-road collections cannot supply.

Core claim

The paper introduces PINNS, a dataset of pedestrian-vehicle interactions in unstructured scenes constructed through a video annotation framework that operates on uncalibrated surveillance cameras; the dataset supplies trajectory data together with scene-level information, spans diverse regions and conditions, follows an established Chinese automation standard, and is accompanied by an analysis showing why such data are needed for progress on heterogeneous trajectory prediction.

What carries the argument

The annotation framework that processes uncalibrated surveillance video to produce trajectory annotations and scene-level information for heterogeneous agents.

Load-bearing premise

Trajectories and interactions can be extracted reliably from uncalibrated surveillance video without systematic errors that would mislead downstream prediction models.

What would settle it

A controlled comparison in which prediction models trained on PINNS show no accuracy gain over models trained only on structured-road datasets when tested on the same unstructured scenes, or an independent audit that measures high rates of annotation mismatch with ground-truth trajectories.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Trajectory prediction research gains access to labeled examples of dense pedestrian-vehicle interactions outside structured roads.
  • Models can be evaluated under variations in weather, lighting, and regional driving norms that current benchmarks omit.
  • The extensible design allows incremental addition of new scenes while maintaining consistent annotation standards.
  • Analysis of current challenges supplies explicit directions for future work on heterogeneous agent prediction.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same framework could be applied to other uncalibrated video sources such as dash cams to enlarge coverage without new hardware.
  • Downstream safety systems for autonomous vehicles might incorporate uncertainty estimates derived from the annotation process itself.
  • Cross-dataset transfer experiments become feasible, testing whether knowledge from structured roads improves or harms performance once unstructured data are added.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 1 minor

Summary. The manuscript introduces PINNS, a dataset and annotation framework for pedestrian-vehicle interactions captured from uncalibrated surveillance cameras in unstructured scenes. It covers multiple countries and regions with diverse traffic scenarios, seasonal/lighting/weather variations, and dense interactions; provides trajectory data plus scene-level information following Chinese Association of Automation standards; analyzes challenges in heterogeneous agent trajectory prediction; and releases the data publicly via GitHub to support research on trajectory prediction and autonomous driving in mixed traffic.

Significance. If the trajectory annotations are shown to be reliable, the dataset would address a documented scarcity of public resources for dense, heterogeneous interactions in unstructured environments and could support improved prediction models for autonomous driving safety. The multi-region coverage, environmental diversity, extensibility design, and public GitHub release are concrete strengths that would aid reproducibility and community extension if the core accuracy claim holds.

major comments (1)
  1. [Annotation Framework / Dataset Construction] The central usefulness claim for downstream prediction models rests on the assertion that trajectories and interactions can be reliably extracted and annotated from uncalibrated video. No quantitative validation of annotation accuracy, error rates, comparison to calibrated ground truth, or analysis of systematic biases from perspective distortion, depth ambiguity, or occlusion appears in the framework description or dataset construction sections, which is load-bearing for the paper's contribution.
minor comments (1)
  1. [Abstract / Title] The acronym construction in the title and abstract (uNcalibrated, uNstructured) is unconventional; a brief note on the intended pronunciation or rationale would improve clarity.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive comments. We address the major point below and will revise the manuscript to strengthen the validation of the annotation framework.

read point-by-point responses

  1. Referee: [Annotation Framework / Dataset Construction] The central usefulness claim for downstream prediction models rests on the assertion that trajectories and interactions can be reliably extracted and annotated from uncalibrated video. No quantitative validation of annotation accuracy, error rates, comparison to calibrated ground truth, or analysis of systematic biases from perspective distortion, depth ambiguity, or occlusion appears in the framework description or dataset construction sections, which is load-bearing for the paper's contribution.

    Authors: We agree that the absence of quantitative validation is a significant gap, as the reliability of trajectories from uncalibrated cameras is central to the dataset's value. The manuscript describes the annotation process and standards followed but does not report error metrics or bias analysis. In the revised version, we will add a new subsection under dataset construction that includes: (1) results from manual verification on randomly sampled trajectories (e.g., pixel-level and world-coordinate consistency checks), (2) discussion of systematic biases arising from perspective distortion, depth ambiguity, and occlusion with qualitative examples, and (3) any available consistency metrics across multiple annotators or temporal frames. We note that calibrated ground-truth data is unavailable by design for these surveillance-camera scenes, so direct comparison is not possible; however, the added analysis will use alternative proxies to quantify reliability. revision: yes

Circularity Check

0 steps flagged

Dataset release paper with no equations, predictions, or self-referential derivations

full rationale

The manuscript presents the PINNS dataset and an annotation framework for trajectories from uncalibrated cameras. No mathematical derivations, fitted parameters, or predictions appear in the provided text. The central claim is the dataset's coverage and extensibility, released via external GitHub link. No steps reduce by construction to inputs, self-citations, or ansatzes; the contribution is the data collection and annotation process itself rather than a claimed derivation chain.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no explicit free parameters, axioms, or invented entities are stated. The framework implicitly assumes that uncalibrated video yields usable trajectory annotations.

pith-pipeline@v0.9.1-grok · 5754 in / 974 out tokens · 21225 ms · 2026-06-29T22:24:53.017662+00:00 · methodology

0 comments
read the original abstract

Predicting the interaction between pedestrian and vehicle is essential for autonomous driving safety in unstructured and semi-structured scenarios; however, this task is severely hindered by the scarcity of public datasets that feature dense pedestrian-vehicle interactions. Most current studies rely on structured road data, leaving the complex, heterogeneous interactions found in unstructured environments insufficiently represented and researched. In this paper, we propose a dataset annotation framework based on video data from uncalibrated surveillance cameras and present PINNS (Pedestrian-vehicle Interaction dataset from uNcalibrated cameras in uNstructured Scenes). The dataset covers multiple countries and regions, includes diverse typical traffic scenarios, and considers variations in seasons, lighting conditions, and weather. It focuses on complex scenes with dense pedestrian-vehicle interactions and is designed to be easily extensible. The dataset is constructed and annotated according to the standard issued by the Chinese Association of Automation, providing both trajectory data and corresponding scene-level information. Furthermore, this paper analyzes current challenges and research directions in heterogeneous agent trajectory prediction, shows the necessity and usefulness of the proposed dataset. We hope our framework and dataset will facilitate research on trajectory prediction and autonomous driving in complex mixed traffic scenarios. PINNS is publicly available at https://github.com/Songan-Lab.

Figures

Figures reproduced from arXiv: 2605.25947 by Haoyang Peng, Ming Yang, Qian Hu, Songan Zhang.

Figure 1
Figure 1. Figure 1: Geographical distribution of dataset scenes. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Reprojection and reconstruction visualization of the homography [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Interaction scene images under different environments. Vertical axis corresponds to different seasons, times of day, and weather conditions, while the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Overview of the dataset construction pipeline. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Comparison of data distributions between the raw data and the filtered dataset. Raw scenes are counted by recording hours, while the filtered dataset [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Distribution of images and annotated targets under different conditions. [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Speed distributions of pedestrians and vehicles in different scenes and regions. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Distributions of the minimum distances between pedestrians and vehicles in different scenes and locations. [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

47 extracted references

  1. [1]

    W. H. Organization,Global Status Report on Road Safety 2023: Sum- mary. World Health Organization, 2023

    2023

  2. [2]

    Motor Vehicle Crash Injury Rates by Mode of Travel, United States: Using Exposure-Based Methods to Quantify Differences,

    L. F. Beck, A. M. Dellinger, and M. E. O’Neil, “Motor Vehicle Crash Injury Rates by Mode of Travel, United States: Using Exposure-Based Methods to Quantify Differences,”American Journal of Epidemiology, vol. 166, no. 2, pp. 212–218, Jun. 2007

    2007

  3. [3]

    Autonomous Vehicles That Interact With Pedestrians: A Survey of Theory and Practice,

    A. Rasouli and J. K. Tsotsos, “Autonomous Vehicles That Interact With Pedestrians: A Survey of Theory and Practice,”IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 3, pp. 900–918, Mar. 2020

    2020

  4. [4]

    Characterizing Structured Versus Unstructured Environments Based on Pedestrians’ and Vehicles’ Motion Trajectories,

    M. Golchoubian, M. Ghafurian, N. L. Azad, and K. Dautenhahn, “Characterizing Structured Versus Unstructured Environments Based on Pedestrians’ and Vehicles’ Motion Trajectories,” in2022 IEEE 25th International Conference on Intelligent Transportation Systems (ITSC). Macau, China: IEEE, Oct. 2022, pp. 2888–2895

    2022

  5. [5]

    You’ll never walk alone: Modeling social behavior for multi-target tracking,

    S. Pellegrini, A. Ess, K. Schindler, and L. Van Gool, “You’ll never walk alone: Modeling social behavior for multi-target tracking,” in2009 IEEE 12th International Conference on Computer Vision. Kyoto: IEEE, Sep. 2009, pp. 261–268

    2009

  6. [6]

    Crowds by Example,

    A. Lerner, Y . Chrysanthou, and D. Lischinski, “Crowds by Example,” Computer Graphics Forum, vol. 26, no. 3, pp. 655–664, Sep. 2007

    2007

  7. [7]

    nuScenes: A Multimodal Dataset for Autonomous Driving,

    H. Caesar, V . Bankiti, A. H. Lang, S. V ora, V . E. Liong, Q. Xu, A. Kr- ishnan, Y . Pan, G. Baldan, and O. Beijbom, “nuScenes: A Multimodal Dataset for Autonomous Driving,” in2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, W A, USA: IEEE, Jun. 2020, pp. 11 618–11 628

    2020

  8. [8]

    Scalability in Per- ception for Autonomous Driving: Waymo Open Dataset,

    P. Sun, H. Kretzschmar, X. Dotiwalla, A. Chouard, V . Patnaik, P. Tsui, J. Guo, Y . Zhou, Y . Chai, B. Caine, V . Vasudevan, W. Han, J. Ngiam, H. Zhao, A. Timofeev, S. Ettinger, M. Krivokon, A. Gao, A. Joshi, Y . Zhang, J. Shlens, Z. Chen, and D. Anguelov, “Scalability in Per- ception for Autonomous Driving: Waymo Open Dataset,” in2020 IEEE/CVF Conference...

    2020

  9. [9]

    Vision meets robotics: The KITTI dataset,

    A. Geiger, P. Lenz, C. Stiller, and R. Urtasun, “Vision meets robotics: The KITTI dataset,”The International Journal of Robotics Research, vol. 32, no. 11, pp. 1231–1237, Sep. 2013

    2013

  10. [10]

    Argoverse: 3D Tracking and Forecasting With Rich Maps,

    M.-F. Chang, D. Ramanan, J. Hays, J. Lambert, P. Sangkloy, J. Singh, S. Bak, A. Hartnett, D. Wang, P. Carr, and S. Lucey, “Argoverse: 3D Tracking and Forecasting With Rich Maps,” in2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA: IEEE, Jun. 2019, pp. 8740–8749

    2019

  11. [11]

    Learning Social Etiquette: Human Trajectory Understanding In Crowded Scenes,

    A. Robicquet, A. Sadeghian, A. Alahi, and S. Savarese, “Learning Social Etiquette: Human Trajectory Understanding In Crowded Scenes,” inComputer Vision – ECCV 2016, B. Leibe, J. Matas, N. Sebe, and M. Welling, Eds. Cham: Springer International Publishing, 2016, vol. 9912, pp. 549–565

    2016

  12. [12]

    The inD Dataset: A Drone Dataset of Naturalistic Road User Tra- jectories at German Intersections,

    J. Bock, R. Krajewski, T. Moers, S. Runde, L. Vater, and L. Eckstein, “The inD Dataset: A Drone Dataset of Naturalistic Road User Tra- jectories at German Intersections,” in2020 IEEE Intelligent Vehicles Symposium (IV). Las Vegas, NV , USA: IEEE, Oct. 2020, pp. 1929– 1934

    2020

  13. [13]

    ParkDiffusion: Heterogeneous Multi-Agent Multi-Modal Trajectory Prediction for Au- tomated Parking using Diffusion Models,

    J. Wei, N. V ¨odisch, A. Rehr, C. Feist, and A. Valada, “ParkDiffusion: Heterogeneous Multi-Agent Multi-Modal Trajectory Prediction for Au- tomated Parking using Diffusion Models,” Aug. 2025

    2025

  14. [14]

    INTERACTION Dataset: An INTERnational, Adversarial and Cooper- ative moTION Dataset in Interactive Driving Scenarios with Semantic Maps,

    W. Zhan, L. Sun, D. Wang, H. Shi, A. Clausse, M. Naumann, J. Kum- merle, H. Konigshof, C. Stiller, A. d. L. Fortelle, and M. Tomizuka, “INTERACTION Dataset: An INTERnational, Adversarial and Cooper- ative moTION Dataset in Interactive Driving Scenarios with Semantic Maps,” Sep. 2019

    2019

  15. [15]

    Top-view Trajectories: A Pedestrian Dataset of Vehicle-Crowd Interaction from Controlled Experiments and Crowded Campus,

    D. Yang, L. Li, K. Redmill, and U. Ozguner, “Top-view Trajectories: A Pedestrian Dataset of Vehicle-Crowd Interaction from Controlled Experiments and Crowded Campus,” in2019 IEEE Intelligent Vehicles Symposium (IV). Paris, France: IEEE, Jun. 2019, pp. 899–904

    2019

  16. [16]

    The ApolloScape Dataset for Autonomous Driving,

    X. Huang, X. Cheng, Q. Geng, B. Cao, D. Zhou, P. Wang, Y . Lin, and R. Yang, “The ApolloScape Dataset for Autonomous Driving,” in2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). Salt Lake City, UT: IEEE, Jun. 2018, pp. 1067– 10 676

    2018

  17. [17]

    One Thousand and One Hours: Self- driving Motion Prediction Dataset,

    J. Houston, G. Zuidhof, L. Bergamini, Y . Ye, L. Chen, A. Jain, S. Omari, V . Iglovikov, and P. Ondruska, “One Thousand and One Hours: Self- driving Motion Prediction Dataset,” Nov. 2020

    2020

  18. [18]

    OnSiteVRU: A High-Resolution Trajectory Dataset for High-Density Vulnerable Road Users,

    Z. Yan, J. Li, P. Hang, and J. Sun, “OnSiteVRU: A High-Resolution Trajectory Dataset for High-Density Vulnerable Road Users,” Mar. 2025

    2025

  19. [19]

    Kinematic 3D Object Detection in Monocular Video,

    G. Brazil, G. Pons-Moll, X. Liu, and B. Schiele, “Kinematic 3D Object Detection in Monocular Video,” inComputer Vision – ECCV 2020, A. Vedaldi, H. Bischof, T. Brox, and J.-M. Frahm, Eds. Cham: Springer International Publishing, 2020, vol. 12368, pp. 135–152

    2020

  20. [20]

    Joint Monocular 3D Vehicle Detection and Tracking,

    H.-N. Hu, Q.-Z. Cai, D. Wang, J. Lin, M. Sun, P. Krahenbuhl, T. Darrell, and F. Yu, “Joint Monocular 3D Vehicle Detection and Tracking,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 5390–5399

    2019

  21. [21]

    Hierarchical camera auto- calibration for traffic surveillance systems,

    S. ´Alvarez, D. Llorca, and M. Sotelo, “Hierarchical camera auto- calibration for traffic surveillance systems,”Expert Systems with Ap- plications, vol. 41, no. 4, pp. 1532–1542, Mar. 2014

    2014

  22. [22]

    Online camera auto-calibration appliable to road surveillance,

    S. Guo, X. Yu, Y . Sha, Y . Ju, M. Zhu, and J. Wang, “Online camera auto-calibration appliable to road surveillance,”Machine Vision and Applications, vol. 35, no. 4, p. 91, Jul. 2024

    2024

  23. [23]

    Camera Auto-Calibration Based on Motion Detection for Airborne Traffic Surveillance,

    M. Qiang and L. Xiling, “Camera Auto-Calibration Based on Motion Detection for Airborne Traffic Surveillance,” in2009 International Conference on Computer Technology and Development. Kota Kinabalu, Malaysia: IEEE, 2009, pp. 350–354

    2009

  24. [24]

    A discrete choice model for solving conflict situations between pedestrians and vehicles in shared space,

    F. Pascucci, N. Rinke, C. Schiermeyer, V . Berkhahn, and B. Friedrich, “A discrete choice model for solving conflict situations between pedestrians and vehicles in shared space,” Sep. 2017

    2017

  25. [25]

    PIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction,

    A. Rasouli, I. Kotseruba, T. Kunic, and J. Tsotsos, “PIE: A Large-Scale Dataset and Models for Pedestrian Intention Estimation and Trajectory Prediction,” in2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seoul, Korea (South): IEEE, Oct. 2019, pp. 6261–6270

    2019

  26. [26]

    Social force model for pedestrian dynamics,

    D. Helbing and P. Moln ´ar, “Social force model for pedestrian dynamics,” Physical Review E: Statistical Physics, Plasmas, Fluids, and Related Interdisciplinary Topics, vol. 51, no. 5, pp. 4282–4286, May 1995

    1995

  27. [27]

    Motion planning in dynamic environments us- ing velocity obstacles,

    P. Fiorini and Z. Shiller, “Motion planning in dynamic environments us- ing velocity obstacles,”The International Journal of Robotics Research, vol. 17, no. 7, pp. 760–772, 1998

    1998

  28. [28]

    Reciprocal Velocity Ob- stacles for real-time multi-agent navigation,

    J. Van Den Berg, Ming Lin, and D. Manocha, “Reciprocal Velocity Ob- stacles for real-time multi-agent navigation,” in2008 IEEE International Conference on Robotics and Automation. Pasadena, CA, USA: IEEE, May 2008, pp. 1928–1935

    2008

  29. [29]

    Simulating dynamical features of escape panic,

    D. Helbing, I. Farkas, and T. Vicsek, “Simulating dynamical features of escape panic,”Nature, vol. 407, no. 6803, pp. 487–490, Sep. 2000

    2000

  30. [30]

    Pedestrian Occupancy Prediction for Autonomous Vehicles,

    P. Zechel, R. Streiter, K. Bogenberger, and U. Gohner, “Pedestrian Occupancy Prediction for Autonomous Vehicles,” in2019 Third IEEE JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 10 International Conference on Robotic Computing (IRC). Naples, Italy: IEEE, Feb. 2019, pp. 230–235

    2021

  31. [31]

    Social LSTM: Human Trajectory Prediction in Crowded Spaces,

    A. Alahi, K. Goel, V . Ramanathan, A. Robicquet, L. Fei-Fei, and S. Savarese, “Social LSTM: Human Trajectory Prediction in Crowded Spaces,” in2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV , USA: IEEE, Jun. 2016, pp. 961– 971

    2016

  32. [32]

    Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks,

    A. Gupta, J. Johnson, L. Fei-Fei, S. Savarese, and A. Alahi, “Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks,” in2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT: IEEE, Jun. 2018, pp. 2255– 2264

    2018

  33. [33]

    SGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction,

    L. Shi, L. Wang, C. Long, S. Zhou, M. Zhou, Z. Niu, and G. Hua, “SGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction,” in2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, TN, USA: IEEE, Jun. 2021, pp. 8990–8999

    2021

  34. [34]

    SocialCircle: Learning the Angle-based Social Interaction Representation for Pedestrian Trajec- tory Prediction,

    C. Wong, B. Xia, Z. Zou, Y . Wang, and X. You, “SocialCircle: Learning the Angle-based Social Interaction Representation for Pedestrian Trajec- tory Prediction,” in2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, W A, USA: IEEE, Jun. 2024, pp. 19 005–19 015

    2024

  35. [35]

    Trajec- tron++: Dynamically-Feasible Trajectory Forecasting With Heteroge- neous Data,

    T. Salzmann, B. Ivanovic, P. Chakravarty, and M. Pavone, “Trajec- tron++: Dynamically-Feasible Trajectory Forecasting With Heteroge- neous Data,” Jan. 2021

    2021

  36. [36]

    TraPHic: Tra- jectory Prediction in Dense and Heterogeneous Traffic Using Weighted Interactions,

    R. Chandra, U. Bhattacharya, A. Bera, and D. Manocha, “TraPHic: Tra- jectory Prediction in Dense and Heterogeneous Traffic Using Weighted Interactions,” in2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach, CA, USA: IEEE, Jun. 2019, pp. 8475–8484

    2019

  37. [37]

    Traf- ficPredict: Trajectory Prediction for Heterogeneous Traffic-Agents,

    Y . Ma, X. Zhu, S. Zhang, R. Yang, W. Wang, and D. Manocha, “Traf- ficPredict: Trajectory Prediction for Heterogeneous Traffic-Agents,”Pro- ceedings of the AAAI Conference on Artificial Intelligence, vol. 33, no. 01, pp. 6120–6127, Jul. 2019

    2019

  38. [38]

    Scene Transformer: A unified architecture for predicting multiple agent trajectories,

    J. Ngiam, B. Caine, V . Vasudevan, Z. Zhang, H.-T. L. Chiang, J. Ling, R. Roelofs, A. Bewley, C. Liu, A. Venugopal, D. Weiss, B. Sapp, Z. Chen, and J. Shlens, “Scene Transformer: A unified architecture for predicting multiple agent trajectories,” Mar. 2022

    2022

  39. [39]

    Query-Centric Trajectory Prediction,

    Z. Zhou, J. Wang, Y .-H. Li, and Y .-K. Huang, “Query-Centric Trajectory Prediction,” in2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, BC, Canada: IEEE, Jun. 2023, pp. 17 863–17 873

    2023

  40. [40]

    SIMPL: A Simple and Efficient Multi-Agent Motion Prediction Baseline for Autonomous Driving,

    L. Zhang, P. Li, S. Liu, and S. Shen, “SIMPL: A Simple and Efficient Multi-Agent Motion Prediction Baseline for Autonomous Driving,” IEEE Robotics and Automation Letters, vol. 9, no. 4, pp. 3767–3774, Apr. 2024

    2024

  41. [41]

    MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction,

    B. Varadarajan, A. Hefny, A. Srivastava, K. S. Refaat, N. Nayakanti, A. Cornman, K. Chen, B. Douillard, C. P. Lam, D. Anguelov, and B. Sapp, “MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction,” in2022 International Conference on Robotics and Automation (ICRA). Philadelphia, PA, USA: IEEE, May 2022, pp. 7814–7821....

    2022

  42. [42]

    (2025–2026) Webcam Live Abetone (PT) - Piazza Europa

    Paesaggi Digitali. (2025–2026) Webcam Live Abetone (PT) - Piazza Europa. [Online]. Available: https://www.youtube.com/watch?v= fjnJ2fzlh4g

    2025

  43. [43]

    (2025–2026) Shinjuku Kabukicho Live Traffic Camera

    KABUKICHO. (2025–2026) Shinjuku Kabukicho Live Traffic Camera. [Online]. Available: https://www.youtube.com/watch?v= EHkMjfMw7oU

    2025

  44. [44]

    (2025–2026) Jackson Hole Town Square Pedestrian Webcam Stream

    See Jackson Hole. (2025–2026) Jackson Hole Town Square Pedestrian Webcam Stream. [Online]. Available: https://www.youtube.com/watch? v=1EiC9bvVGnk

    2025

  45. [45]

    (2025–2026) Live Camera Stream of Chaweng, Koh Samui [Thailand]

    The Real Samui Webcam. (2025–2026) Live Camera Stream of Chaweng, Koh Samui [Thailand]. [Online]. Available: https: //www.youtube.com/watch?v=VR-x3HdhKLQ

    2025

  46. [46]

    ByteTrack: Multi-object tracking by associating every detection box,

    Y . Zhang, P. Sun, Y . Jiang, D. Yu, F. Weng, Z. Yuan, P. Luo, W. Liu, and X. Wang, “ByteTrack: Multi-object tracking by associating every detection box,” inComputer Vision – ECCV 2022, S. Avidan, G. Bros- tow, M. Ciss ´e, G. M. Farinella, and T. Hassner, Eds. Cham: Springer Nature Switzerland, 2022, pp. 1–21

    2022

  47. [47]

    Monocular 3D Vehicle Detection Using Uncalibrated Traffic Cameras through Homography,

    M. Zhu, S. Zhang, Y . Zhong, P. Lu, H. Peng, and J. Lenneman, “Monocular 3D Vehicle Detection Using Uncalibrated Traffic Cameras through Homography,” in2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Prague, Czech Republic: IEEE, Sep. 2021, pp. 3814–3821. VI. BIOGRAPHY Haoyang Pengreceived the B.S. degree in Artificial Int...

    2021