The DOI matches its Crossref record; 55 of the 64 references it cites were found in a registry or an index.
SceneGen: Learning to Generate Realistic Traffic Scenes
Open accessAbstract
We consider the problem of generating realistic traffic scenes automatically. Existing methods typically insert actors into the scene according to a set of hand-crafted heuristics and are limited in their ability to model the true complexity and diversity of real traffic scenes, thus inducing a content gap between synthesized traffic scenes versus real ones. As a result, existing simulators lack the fidelity necessary to train and test self-driving vehicles. To address this limitation, we present SceneGen—a neural autoregressive model of traffic scenes that eschews the need for rules and heuristics. In particular, given the ego-vehicle state and a high definition map of surrounding area, SceneGen inserts actors of various classes into the scene and synthesizes their sizes, orientations, and velocities. We demonstrate on two large-scale datasets SceneGen’s ability to faithfully model distributions of real traffic scenes. Moreover, we show that SceneGen coupled with sensor simulation can be used to train perception models that generalize to the real world.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Generating large synthetic datasets has been a problem of great interest in other domains as well [71-77], and has also been used as a tool for benchmarking motion planning [78].”
From MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations (2023) · cites this paper as [77]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Hochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“The basis of our model is the ConvLSTM architecture [42]—an extension of the classic LSTM architecture [22] to spatial data—and the input to our model at the i-th generation step is a bird’s eye view multi-channel image encoding the SDV a0, the HD map m, and the actors generated so far {a1, .”
MethodAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Book Review: Pattern Recognition and Machine Learning
Nasrabadi, N. M. (2007) · Journal of Electronic Imaging · doi:10.1117/1.2819119 · Cited by 22,066 · cited 1×
“Each actor’s bounding box is sampled from a Gaussian KDE fitted to the training dataset [5] and velocities are set to satisfy speed limits and a time gap between successive actors on the lane graph.”
MethodGANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium
Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., & Hochreiter, S. (2017) · arXiv · doi:10.48550/arxiv.1706.08500 · Cited by 8,365
Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
Shi, X., Chen, Z., Wang, H., Yeung, D.-Y., Wong, W.-K., & Woo, W.-C. (2015) · arXiv · doi:10.48550/arxiv.1506.04214 · Cited by 7,709
Improved Techniques for Training GANs
Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., & Chen, X. (2016) · arXiv · doi:10.48550/arxiv.1606.03498 · Cited by 6,965 · cited 1×
“This is akin to the popular IS [41], FID [21], and KID [4] metrics for evaluating generative models, except we use a feature extractor trained on traffic scenes.”
MethodWaveNet: a generative model for raw audio
Oord, A. V. D., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016) · arXiv · doi:10.48550/arxiv.1609.03499 · Cited by 4,911 · cited 1×
“Recently, neural autoregressive models have found tremendous success in modeling a variety of data, including handwriting [18], images [49], audio [48], text [39], sketches [20], graphs [29], 3D meshes [33], indoor scenes [50] and image scene layouts [25].”
BackgroundCongested traffic states in empirical observations and microscopic simulations
Treiber, M., Hennecke, A., & Helbing, D. (2000) · Physical Review E · doi:10.1103/physreve.62.1805 · Cited by 4,762 · cited 1×
“In contrast, microscopic models simulate the behavior of each individual vehicle over time by assuming a car-following model [36, 6, 34, 13, 17, 1, 44].”
BackgroundOn kinematic waves II. A theory of traffic flow on long crowded roads
Lighthill, S. J., & Whitham, G. B. (1955) · Proceedings of the Royal Society of London A Mathematical and Physical Sciences · doi:10.1098/rspa.1955.0089 · Cited by 4,606
Microscopic Traffic Simulation using SUMO
López, P. Á., Wießner, E., Behrisch, M., Bieker-Walz, L., Erdmann, J., Flötteröd, Y.-P., Hilbrich, R., Lücken, L., Rummel, J., & Wagner, P. (2018) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc.2018.8569938 · Cited by 3,898 · cited 7×
“Microscopic traffic models have been included in popular software packages such as SUMO [31], CORSIM [35], VISSIM [11], and MITSIM [55].”
BackgroundMethodInfluential
Flagged first, then by citations
9 of the 64 references couldn’t be matched to a known paper.
Research works printed but not linked: 3
- gerlough, 1955, Simulation of Freeway Traffic on a General-Purpose Discrete Variable Computer
- uria, 2013, NADE: the real-valued neural autoregressive density-estimator, CoRR
- taheri, 1990, An investigation and design of slip control braking systems integrated with four wheel steering
Cited by
Sorted by citations
Items 101–105 of 105
Long-Term Traffic Simulation via Structured Autoregressive Modeling
Xiao, L., Feng, Z., & Yan, X. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37550-6_37
ProtoAug: Prototype-Guided Uncertainty-Aware Augmentation for Long-Tail Motion Prediction
Lu, Z., Cai, Y., Wang, H., Chen, L., Wang, Y., & Zuo, X. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3710064
“Beyond target-agent trajectory synthesis, SceneGen [40] explores full-scene tra ffi c generation for realistic simulation, while Tra ff ormer [41] models long-horizon spatiotemporal tra ffi c states over road networks, providing complementary perspectives on tra ffi c-level generation and forecasting.”
Yang, K.-X., & Qu, S.-R. (2026) · Italian National Conference on Sensors · doi:10.3390/s26154869
“Recent studies have developed various learning-based generation frameworks, including traffic scene generation, unified initial-state-and-trajectory modeling, controllable diffusion-based generation, and language-conditioned traffic generation [1–5].”
Li, X., Hu, X., Wang, L., et al. (2026) · Advanced Engineering Informatics · doi:10.1016/j.aei.2026.105194
MDG: Masked Denoising Generation for Multi-Agent Behavior Modeling in Traffic Environments
Huang, Z., Zhou, Z., Cai, T., Zhang, Y., & Ma, J. (2026) · doi:10.1109/cvprf71909.2026.00089
“Recent advances in generative modeling have improved realism and diversity by learning joint scene-level distributions over agent behaviors [43, 44, 55].”
Metadata
- Authors
- Tan, S., Wong, K., Wang, S., Manivasagam, S., Ren, M., & Urtasun, R.
- Affiliations
- Sun Yat-sen University (CN)
- University of Toronto (CA)
- Venue
- 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2021-06
- Pages
- 892-901
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W3173280621
- Semantic Scholar
- 8b817a28d880e4688956d6e1769475a9f22c7411
- Cited by
- 104
- References
- 64
- Citations by year
- 2026: 10 · 2025: 36 · 2024: 29 · 2023: 17 · 2022: 8 · 2021: 4
- FWCI
- 14.34
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion, cc-by)
- arXiv (Cornell University) (cc-by)
- Topics
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Physical Sciences › Engineering › Control and Systems Engineering › Simulation and Modeling Applications
- Keywords
- autonomous driving simulation, high-definition maps, sensor simulation
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
