Pith. sign in

REVIEW 3 major objections 5 minor 71 references

SynSHRP2: A Synthetic Multimodal Benchmark for Driving Safety-critical Events Derived from Real-world Driving Data

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper introduces SynSHRP2, a public synthetic dataset that re-renders 1,874 crashes and 6,924 near-crashes from SHRP 2 with personally identifiable content removed while preserving the safety-critical scene.

desk verdict Potentially valuable public synthetic SCE dataset from SHRP2, but the central de-identification and fidelity claims need quantitative validation before the dataset can be trusted. read the letter →

arxiv 2505.06276 v1 pith:INHV3DEV submitted 2025-05-06 cs.RO

Liang Shi , Boyu Jiang , Zhenyuan Yuan , Miguel A. Perez , Feng Guo This is my paper · ORCID
classification cs.RO
keywords syntheticdrivingdatasetsafety-criticaleventsSHRP2naturalisticstudyde-identificationStableDiffusionControlNetbenchmarkmultimodaldata
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that the largest naturalistic driving study's safety-critical events can be made publicly shareable by re-synthesizing each frame with a diffusion model that strips out faces, license plates, and other personal identifiers while keeping the geometry and timing that matter for crash analysis. To back this, it builds SynSHRP2: a multimodal dataset of 1,874 crashes and 6,924 near-crashes, each with five de-identified synthetic keyframes, kinematic time series spanning five seconds around the event, tabular annotations, and hand-written narratives. If the de-identification truly preserves the safety-relevant content, this would give any researcher open access to thousands of rare events that are currently locked behind certification and data-use agreements.

What carries the argument

The central object is the per-object 'segmentation-masked, orientation-conditioned synthesis' workflow: semantic segmentation classifies every pixel region, objects carrying PII are re-generated by Stable Diffusion with two ControlNets—one fed the Canny edge sketch to lock the object's facing direction, one fed the original object segment to fix its position—and IP-Adapter uses the Impact keyframe as an image prompt to keep the same vehicle or pedestrian appearance across the other keyframes. This combination is what is supposed to let the dataset delete personal data without deleting the crash-relevant information.

What would settle it

Take a random sample of SynSHRP2 events with their original SHRP 2 keyframes, and measure per-object agreement—bounding-box IoU for vehicles and pedestrians, orientation classification (front vs. rear), lane occupancy, and brake-light state—between the synthetic keyframe and the original; if median IoU falls below about 0.5, or if more than a few percent of objects flip orientation, or if a face-recognition or vehicle-re-identification model can match synthetic frames to original drivers, the core claim of preservation-with-de-identification is refuted.

Watch

Extended reading notes

Core claim

The central claim is that a diffusion-based per-object synthesis pipeline can convert restricted SHRP 2 naturalistic driving videos into a public dataset in which personally identifiable information is removed while the elements that define a safety-critical event are preserved: vehicle locations, orientations, road geometry, traffic-control devices, and the temporal sequence across five keyframes. The pipeline upscales each keyframe, segments it semantically, re-synthesizes only the PII-bearing objects under two ControlNets (Canny edges for orientation, the original segment for location), and uses IP-Adapter anchored on the Impact frame to keep each object's appearance consistent across frames. The paper argues these components together ensure that the synthetic frames retain the spatial-temporal relationships among road users and the scene setup that a crash analyst or an automated-driving system would need.

Load-bearing premise

The load-bearing premise is that each object re-synthesized by Stable Diffusion and ControlNet stays at the same location, with the same orientation and interaction geometry, across all five keyframes, and that IP-Adapter keeps objects consistent between frames, so the generated images faithfully represent the original safety-critical event while removing all personally identifiable information.

Editorial extensions

If this is right

  • Any researcher can now train and fairly compare SCE attribute classifiers on one fixed public benchmark with standard splits, removing the configuration differences that made prior SHRP 2 results hard to compare.
  • Vision-language models can be evaluated on safety-critical scene understanding against ground-truth narratives without ever viewing a real driver's face or license plate.
  • The five-second kinematic window around each event, paired with the synthetic keyframes, supports multimodal crash prediction and driver-response studies in the open.
  • The published benchmarks (six kinematic classifiers and six vision-language models) establish baselines against which future work on this dataset can be measured.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the sensor data is unmodified original SHRP 2 data, the attribute-detection benchmarks actually measure the kinematics of the real events, not the fidelity of the synthetic images; a cleaner test of image fidelity would compare downstream performance on original versus synthetic keyframes for the same events.
  • The paper does not run a quantitative fidelity or privacy audit, so a natural extension is measuring per-object agreement (bounding-box overlap, orientation, lane position) between synthetic and original keyframes on a sample of events, and testing re-identification attacks on the synthetic frames.
  • The authors note future work on synthesizing full de-identified video; extending from five keyframes to continuous video would stress the IP-Adapter consistency mechanism much harder, since there would be no single Impact anchor for long sequences.
  • If the de-identification is demonstrated to be strong, this release model could become a template for sharing other restricted naturalistic driving datasets, lowering the privacy barrier for safety research worldwide.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces SynSHRP2, a synthetic multimodal driving dataset derived from the SHRP 2 NDS, containing 1,874 crashes and 6,924 near-crash events. Each event includes tabular attributes, 5-second time-series kinematic data, five keyframe images, and annotated narrative descriptions. The keyframes are de-identified through a Stable Diffusion/ControlNet pipeline that uses semantic segmentation, Canny edge constraints, and IP-Adapter for cross-frame consistency. The paper also presents two benchmarks: Task 1 classifies event severity, incident type, and conflict type from kinematic signals with six baseline models; Task 2 evaluates six vision-language models on generating narrative descriptions of the synthetic keyframes. The central claims are that the de-identification preserves critical safety-related information while eliminating personally identifiable information, and that the released dataset provides a useful public benchmark for driving-safety research.

Significance. If the fidelity and privacy claims are substantiated, SynSHRP2 would be a valuable public resource: it offers thousands of real-world-derived safety-critical events with multimodal annotations, directly addressing the rarity and access barriers that limit SHRP 2 NDS use. The paper also provides a clearly documented synthesis pipeline, reproducible benchmark protocols, and comparisons to prior work. However, the load-bearing contribution is the synthetic keyframe dataset, and the current evidence for its two fundamental properties—geometric/semantic fidelity to the original events and complete removal of PII—is qualitative only. The two benchmarks do not test either property, because Task 1 uses only time-series kinematics and Task 2 measures narrative text similarity. For this reason, the paper's central claims require additional validation before the dataset can be used as advertised.

major comments (3)
  1. [Abstract; Section 3.2, Eq. (3)] The claim that the synthetic keyframes 'ensure the preservation of critical safety-related information while eliminating personally identifiable data' is supported only by the qualitative examples in Figures 3-5. The generation step in Eq. (3) constrains the output through Canny edges, semantic segmentation masks, and text prompts, but these do not fix object category, exact bounding-box position, size, count, or fine-grained appearance; a prompt with the semantic class 'vehicle' does not prevent a truck from being replaced by a car, and edge-preserving synthesis can still shift an object's location. No quantitative fidelity metrics are reported, such as mask or bounding-box IoU, category consistency, orientation agreement, or human judgment of safety-relevant scene elements. The two benchmarks in Section 5 do not close this gap: Task 1 uses only kinematic time series, and Task 2 compares narrative text similarity, neither of which verifies geometric fidelity of the synthesized images. Please add a fidelity audit on a representative sample of events with per-object quantitative measures and report failure rates by object category and event type.
  2. [Section 3.2, 'Synthesis of de-identified keyframes'] The privacy claim is not audited. The pipeline explicitly copies non-PII segments unchanged: 'if object O contains no PII, the corresponding segment is directly passed to the blue synthesized frame, i.e., \tilde P_O = P_O.' Consequently, any face, license plate, street sign, or other identifier that semantic segmentation fails to detect remains in the publicly released data. The manuscript reports no face-detection results, no license-plate recognition results, no OCR-based street-name check, and no re-identification risk assessment. A quantitative privacy evaluation, including segmentation false-negative rates on PII classes and manual inspection of a sample, is necessary to substantiate the 'eliminating personally identifiable data' claim.
  3. [Section 3.2 and Section 4.1 (IP-Adapter)] The objective of cross-frame consistency is asserted in Section 3.2 and demonstrated only with the qualitative comparison in Figure 5. Since each event contains five keyframes and the dataset is intended to support scene understanding, the paper should report quantitative consistency measures—for example, object re-identification accuracy across keyframes, appearance-feature distances for matched objects, or temporal bounding-box continuity—and state how many events were processed with the IP-Adapter configuration described. Without such measures, the reproducibility and reliability of the multi-frame synthesis are not established.
minor comments (5)
  1. [Section 3.1 vs. Section 4.1] Section 3.1 lists five critical timestamps (5 s before Event Start, Event Start, Reaction Start, Impact, Event End), but Section 4.1 says 'the remaining five keyframes are generated' after synthesizing the Impact frame, implying six keyframes. Please correct the count and the wording.
  2. [Section 3.2, Eq. (3)] The notation around Eq. (3) is unclear: 'SD(ConNet(LO, TO), ConNet(PO); z)' mixes a function with side inputs, and the statement that the synthesized segment and original segment 'should have the same number of pixels' should be reconciled with mask resizing and the super-resolution step. A precise definition of how masks are aligned to the generated segment would help.
  3. [Section 3.2 text] The phrase 'directly passed to the blue synthesized frame' appears to be a typo; 'blended' or 'final' synthesized frame is likely intended.
  4. [Table 1 heading] The heading 'Macro Precsion' contains a typo; it should read 'Macro Precision.'
  5. [Dataset release information] The abstract and introduction state that SynSHRP2 is publicly available, but the manuscript provides no URL, license, access procedure, or data-card information. Please include the release location and usage terms.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the dataset derives from external SHRP 2 NDS data and the benchmarks use held-out test splits and manually annotated targets.

full rationale

Checked the claimed derivation chain for the defined circularity patterns; none is exhibited. SynSHRP2's content is derived from the external SHRP 2 NDS ('over 1,874 crashes and 6,924 near-crashes derived from the SHRP 2 NDS'), not from a quantity that the paper itself fits and then 'predicts'. The de-identification/synthesis pipeline (Section 3.2, Eq. 1-3) is a generation procedure using StableSR, semantic segmentation, Canny ControlNet, and IP-Adapter; it does not present a fitted parameter renamed as a prediction, and the 'ensuring' language in the abstract is an empirical fidelity/privacy claim (supported only qualitatively in Figures 3-5), not a derivation that reduces to its inputs. Benchmark Task 1 (Section 5.1) states 'The dataset was randomly divided into training, testing, and validation subsets in the proportion of 7:2:1' and reports on an 'independent testing set', so the reported accuracy numbers are not forced by construction. Task 2 compares VLM narratives against manually coded ground-truth narratives; even though the prompt injects event severity and conflict type, the metric is narrative similarity, not a predicted quantity equal to the injected labels, so any leakage is a benchmark-validity concern rather than a circular derivation. Self-citations occur (Shi et al. [46], [47] for benchmark models; Dingus et al. [14] and Hankey et al. [22] for SHRP 2 context), but they are not the load-bearing justification of the dataset's central premise, which rests on the external SHRP 2 data and independent test splits. I therefore find no significant circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the reliability of SHRP 2 annotations and on the unvalidated assumption that the generative pipeline preserves safety-critical information while removing PII. No free parameters are fitted in the paper; the benchmark hyperparameters are standard and not used to support the central claim.

assumptions (3)
  • domain assumption The SHRP 2 NDS event annotations (event type, conflict type, incident type, crash severity, narratives) are accurate and reliable.
    The dataset and all benchmark ground truths inherit the quality of SHRP 2 annotations. They are treated as ground truth without independent verification. Used throughout Sections 3.1 and 5.
  • ad hoc to paper The Stable Diffusion/ControlNet de-identification pipeline preserves the spatial relationships, object orientations, and traffic context of each original keyframe while removing PII.
    This is the load-bearing assumption for the dataset's utility and privacy claims. It is only demonstrated qualitatively in Figures 3-5, not validated quantitatively over the dataset. See Section 3.2.
  • domain assumption The generated keyframes remain temporally aligned and semantically consistent with the time-series sensor data and narrative annotations.
    The multimodal integration assumes that the synthetic images correspond to the same events as the sensor data, and that IP-Adapter preserves cross-frame consistency. Stated in Sections 3.1 and 3.2, not verified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SynSHRP2: A Synthetic Multimodal Benchmark for Driving Safety-critical Events Derived from Real-world Driving Data." pith.science (2026). https://pith.science/paper/INHV3DEV

@misc{pith2026250506276,
  author       = {Pith},
  title        = {Pith review of: SynSHRP2: A Synthetic Multimodal Benchmark for Driving Safety-critical Events Derived from Real-world Driving Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/INHV3DEV}},
  note         = {Machine review of arXiv:2505.06276}
}
read the original abstract

Driving-related safety-critical events (SCEs), including crashes and near-crashes, provide essential insights for the development and safety evaluation of automated driving systems. However, two major challenges limit their accessibility: the rarity of SCEs and the presence of sensitive privacy information in the data. The Second Strategic Highway Research Program (SHRP 2) Naturalistic Driving Study (NDS), the largest NDS to date, collected millions of hours of multimodal, high-resolution, high-frequency driving data from thousands of participants, capturing thousands of SCEs. While this dataset is invaluable for safety research, privacy concerns and data use restrictions significantly limit public access to the raw data. To address these challenges, we introduce SynSHRP2, a publicly available, synthetic, multimodal driving dataset containing over 1874 crashes and 6924 near-crashes derived from the SHRP 2 NDS. The dataset features de-identified keyframes generated using Stable Diffusion and ControlNet, ensuring the preservation of critical safety-related information while eliminating personally identifiable data. Additionally, SynSHRP2 includes detailed annotations on SCE type, environmental and traffic conditions, and time-series kinematic data spanning 5 seconds before and during each event. Synchronized keyframes and narrative descriptions further enhance its usability. This paper presents two benchmarks for event attribute classification and scene understanding, demonstrating the potential applications of SynSHRP2 in advancing safety research and automated driving system development.

Figures

Figures reproduced from arXiv: 2505.06276 by the authors.

Figure 1
Figure 1. Example illustrating all data types in SynSHRP2. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Workflow of the de-identified keyframe synthesis process. The pipeline begins with a keyframe input, applying super-resolution [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Comparison of synthetic image generation approaches. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: Comparison between synthetic images with and without [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

71 extracted references · 49 canonical work pages

  1. [1]

    Llama 3.2: Revolutionizing edge ai and vision with open, customizable models, 2024

    Meta AI. Llama 3.2: Revolutionizing edge ai and vision with open, customizable models, 2024. Accessed: 2025-02-21. 8

    work page 2024

  2. [2]

    Safety crit- ical event prediction through unified analysis of driver and vehicle volatilities: Application of deep learning methods

    Ramin Arvin, Asad J Khattak, and Hairong Qi. Safety crit- ical event prediction through unified analysis of driver and vehicle volatilities: Application of deep learning methods. Accident Analysis & Prevention, 151:105949, 2021. 2, 6, 7

    work page 2021

  3. [3]

    Meteor: An automatic metric for mt evaluation with improved correlation with hu- man judgments

    Satanjeev Banerjee and Alon Lavie. Meteor: An automatic metric for mt evaluation with improved correlation with hu- man judgments. In Proceedings of the acl workshop on in- trinsic and extrinsic evaluation measures for machine trans- lation and/or summarization, pages 65–72, 2005. 8

    work page 2005

  4. [4]

    Uncertainty-based traffic accident anticipation with spatio-temporal relational learn- ing

    Wentao Bao, Qi Yu, and Yu Kong. Uncertainty-based traffic accident anticipation with spatio-temporal relational learn- ing. In Proceedings of the 28th ACM International Confer- ence on Multimedia, pages 2682–2690, 2020. 2

    work page 2020

  5. [5]

    nuscenes: A multi- modal dataset for autonomous driving

    Holger Caesar, Varun Bankiti, Alex H Lang, Sourabh V ora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Gi- ancarlo Baldan, and Oscar Beijbom. nuscenes: A multi- modal dataset for autonomous driving. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11621–11631, 2020. 1, 2

    2020

  6. [6]

    Motion planning diffusion: Learning and planning of robot motions with diffusion models

    Joao Carvalho, An T Le, Mark Baierl, Dorothea Koert, and Jan Peters. Motion planning diffusion: Learning and planning of robot motions with diffusion models. In 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 1916–1923. IEEE, 2023. 3

    work page 2023

  7. [7]

    Anticipating accidents in dashcam videos

    Fu-Hsiang Chan, Yu-Ting Chen, Yu Xiang, and Min Sun. Anticipating accidents in dashcam videos. In Computer Vision–ACCV 2016: 13th Asian Conference on Computer Vision, Taipei, Taiwan, November 20-24, 2016, Revised Se- lected Papers, Part IV 13 , pages 136–153. Springer, 2017. 2

    work page 2016

  8. [8]

    Comfyui: A powerful and modular stable diffusion gui and backend, 2025

    comfyanonymous. Comfyui: A powerful and modular stable diffusion gui and backend, 2025. Accessed: 2025-02-22. 5

    work page 2025

Show all 71 references
  1. [9]

    Xtuner: A toolkit for efficiently fine-tuning llm

    XTuner Contributors. Xtuner: A toolkit for efficiently fine-tuning llm. https://github.com/InternLM/ xtuner, 2023. 7, 8

    2023

  2. [10]

    Diffusion models in vision: A survey

    Florinel-Alin Croitoru, Vlad Hondru, Radu Tudor Ionescu, and Mubarak Shah. Diffusion models in vision: A survey. IEEE Transactions on Pattern Analysis and Machine Intelli- gence, 45(9):10850–10869, 2023. 3 8

    2023

  3. [11]

    Comfyui ipadapter plus, 2025

    cubiq. Comfyui ipadapter plus, 2025. Accessed: 2025-02-

    2025

  4. [12]

    Kitti-carla: a kitti-like dataset generated by carla simulator

    Jean-Emmanuel Deschaud. Kitti-carla: a kitti-like dataset generated by carla simulator. arXiv preprint arXiv:2109.00892, 2021. 2

    2021 arXiv

  5. [13]

    Paris-carla-3d: A real and synthetic outdoor point cloud dataset for challenging tasks in 3d mapping

    Jean-Emmanuel Deschaud, David Duque, Jean Pierre Richa, Santiago Velasco-Forero, Beatriz Marcotegui, and Franc ¸ois Goulette. Paris-carla-3d: A real and synthetic outdoor point cloud dataset for challenging tasks in 3d mapping. Remote Sensing, 13(22):4713, 2021. 2

    2021

  6. [14]

    Dingus, Feng Guo, Suzie Lee, Jonathan F

    Thomas A. Dingus, Feng Guo, Suzie Lee, Jonathan F. Antin, Miguel Perez, Mindy Buchanan-King, and Jonathan Hankey. Driver crash risk factors and prevalence evaluation using nat- uralistic driving data. Proceedings of the National Academy of Sciences, 113(10):2636–2641, 2016. 1

    2016

  7. [15]

    Carla: An open urban driv- ing simulator

    Alexey Dosovitskiy, German Ros, Felipe Codevilla, Anto- nio Lopez, and Vladlen Koltun. Carla: An open urban driv- ing simulator. In Conference on robot learning, pages 1–16. PMLR, 2017. 2

    2017

  8. [16]

    Comfyui-controlnet-aux, 2025

    Fannovel16. Comfyui-controlnet-aux, 2025. Accessed: 2025-02-22. 5, 7

    2025

  9. [17]

    SHRP2 Annual Report

    Federal Highway Administration. SHRP2 Annual Report

  10. [18]

    Vision meets robotics: The kitti dataset

    Andreas Geiger, Philip Lenz, Christoph Stiller, and Raquel Urtasun. Vision meets robotics: The kitti dataset. Interna- tional Journal of Robotics Research (IJRR), 2013. 1, 2

    2013

  11. [19]

    Deep learning, 2016

    Ian Goodfellow. Deep learning, 2016. 2

    2016

  12. [20]

    A review of semantic segmentation using deep neural networks

    Yanming Guo, Yu Liu, Theodoros Georgiou, and Michael S Lew. A review of semantic segmentation using deep neural networks. International journal of multimedia information retrieval, 7:87–93, 2018. 4

    2018

  13. [21]

    Carla-loc: synthetic slam dataset with full-stack sensor setup in chal- lenging weather and dynamic environments

    Yuhang Han, Zhengtao Liu, Shuo Sun, Dongen Li, Jiawei Sun, Chengran Yuan, and Marcelo H Ang Jr. Carla-loc: synthetic slam dataset with full-stack sensor setup in chal- lenging weather and dynamic environments. arXiv preprint arXiv:2309.08909, 2023. 2

    2023 arXiv

  14. [22]

    Description of the shrp 2 naturalistic database and the crash, near-crash, and baseline data sets

    Jonathan M Hankey, Miguel A Perez, and Julie A McClaf- ferty. Description of the shrp 2 naturalistic database and the crash, near-crash, and baseline data sets. Technical report, Virginia Tech Transportation Institute, 2016. 2

    2016

  15. [23]

    Gaia-1: A generative world model for au- tonomous driving

    Anthony Hu, Lloyd Russell, Hudson Yeo, Zak Murez, George Fedoseev, Alex Kendall, Jamie Shotton, and Gian- luca Corrado. Gaia-1: A generative world model for au- tonomous driving. arXiv preprint arXiv:2309.17080, 2023. 2

    2023 arXiv

  16. [24]

    Brain4cars: Car that knows before you do via sensory-fusion deep learning archi- tecture

    Ashesh Jain, Hema S Koppula, Shane Soh, Bharad Ragha- van, Avi Singh, and Ashutosh Saxena. Brain4cars: Car that knows before you do via sensory-fusion deep learning archi- tecture. arXiv preprint arXiv:1601.00740, 2016. 1, 2

    2016 arXiv

  17. [25]

    Real-archvis-xl-xlv10-sdxl, 2025

    John6666. Real-archvis-xl-xlv10-sdxl, 2025. Accessed: 2025-02-22. 5

    2025

  18. [26]

    Anatomically-controllable medical image gen- eration with segmentation-guided diffusion models

    Nicholas Konz, Yuwen Chen, Haoyu Dong, and Maciej A Mazurowski. Anatomically-controllable medical image gen- eration with segmentation-guided diffusion models. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 88–98. Springer,

  19. [27]

    Sinddm: A single image denoising diffusion model

    Vladimir Kulikov, Shahar Yadin, Matan Kleiner, and Tomer Michaeli. Sinddm: A single image denoising diffusion model. In International conference on machine learning , pages 17920–17930. PMLR, 2023. 3

    2023

  20. [28]

    Attention r-cnn for accident detection

    Trung-Nghia Le, Shintaro Ono, Akihiro Sugimoto, and Hi- roshi Kawasaki. Attention r-cnn for accident detection. In 2020 IEEE intelligent vehicles symposium (IV) , pages 313–

    2020

  21. [29]

    Compose and conquer: diffusion-based 3d depth aware composable image synthe- sis

    Jonghyun Lee, Hansam Cho, Youngjoon Yoo, Seoung Bum Kim, and Yonghyun Jeong. Compose and conquer: diffusion-based 3d depth aware composable image synthe- sis. arXiv preprint arXiv:2401.09048, 2024. 3

    2024 arXiv

  22. [30]

    Stabilizing diffusion model for robotic control with dynamic programming and transition feasibility

    Haoran Li, Yaocheng Zhang, Haowei Wen, Yuanheng Zhu, and Dongbin Zhao. Stabilizing diffusion model for robotic control with dynamic programming and transition feasibility. IEEE Transactions on Artificial Intelligence, 2024. 3

    2024

  23. [31]

    Rouge: A package for automatic evaluation of summaries

    Chin-Yew Lin. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out , pages 74–81, 2004. 8

    2004

  24. [32]

    Diff- bir: Toward blind image restoration with generative diffusion prior

    Xinqi Lin, Jingwen He, Ziyan Chen, Zhaoyang Lyu, Bo Dai, Fanghua Yu, Yu Qiao, Wanli Ouyang, and Chao Dong. Diff- bir: Toward blind image restoration with generative diffusion prior. In European Conference on Computer Vision , pages 430–448. Springer, 2024. 3

    2024

  25. [33]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. In NeurIPS, 2023. 8

    2023

  26. [34]

    Diffusion models, image super-resolution, and everything: A survey

    Brian B Moser, Arundhati S Shanbhag, Federico Raue, Stanislav Frolov, Sebastian Palacio, and Andreas Dengel. Diffusion models, image super-resolution, and everything: A survey. IEEE Transactions on Neural Networks and Learn- ing Systems, 2024. 3

    2024

  27. [35]

    Prediction of near-crashes from observed vehicle kinematics using machine learning.Transportation Research Record, 2673(12):463–473, 2019

    Osama A Osman, Mustafa Hajij, Peter R Bakhit, and Sherif Ishak. Prediction of near-crashes from observed vehicle kinematics using machine learning.Transportation Research Record, 2673(12):463–473, 2019. 2, 6, 7

    2019

  28. [36]

    Bleu: a method for automatic evaluation of machine translation

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages 311–318,

  29. [37]

    Driving maneuver early detection via sequence learning from vehicle signals and video images

    Xishuai Peng, Yi Lu Murphey, Ruirui Liu, and Yuanxiang Li. Driving maneuver early detection via sequence learning from vehicle signals and video images. Pattern Recognition, 103:107276, 2020. 2

    2020

  30. [38]

    Transportation safety meets big data: the shrp 2 naturalistic driving database

    Miguel Perez, Shane MCLAUGHLIN, Takayuki Kondo, Jonathan Antin, Julie McClafferty, Suzanne Lee, Jonathan Hankey, and Thomas Dingus. Transportation safety meets big data: the shrp 2 naturalistic driving database. Journal of the Society of Instrument and Control Engineers , 55(5...

    2016

  31. [39]

    Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. Sdxl: Improving latent diffusion mod- els for high-resolution image synthesis. arXiv preprint arXiv:2307.01952, 2023. 5

    2023 arXiv

  32. [40]

    Car crash detection in videos

    Veronica Radu, Mihai Nan, Mihai Tr ˘asc˘au, David Traian Iancu, Alexandra S ¸tefania Ghit ¸˘a, and Adina Magda Flo- 9 rea. Car crash detection in videos. In 2021 23rd Interna- tional Conference on Control Systems and Computer Science (CSCS), pages 127–132. IEEE, 2021. 2

    2021

  33. [41]

    Hanoona Rasheed, Muhammad Maaz, Salman Khan, and Fa- had S. Khan. Llava++: Extending visual capabilities with llama-3 and phi-3, 2024. 8

    2024

  34. [42]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 2, 4

    2022

  35. [43]

    An improved canny edge detection algorithm

    Weibin Rong, Zhanjing Li, Wei Zhang, and Lining Sun. An improved canny edge detection algorithm. In 2014 IEEE international conference on mechatronics and automation , pages 577–582. IEEE, 2014. 4, 5

    2014

  36. [44]

    Benchmarks for retrospective automated driving system crash rate analysis using police-reported crash data

    John M Scanlon, Kristofer D Kusano, Laura A Fraade- Blanar, Timothy L McMurry, Yin-Hsiu Chen, and Trent Vic- tor. Benchmarks for retrospective automated driving system crash rate analysis using police-reported crash data. Traffic Injury Prevention, 25(sup1):S51–S65, 2024. 1

    2024

  37. [45]

    Difftalk: Crafting diffusion models for generalized audio-driven portraits animation

    Shuai Shen, Wenliang Zhao, Zibin Meng, Wanhua Li, Zheng Zhu, Jie Zhou, and Jiwen Lu. Difftalk: Crafting diffusion models for generalized audio-driven portraits animation. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 1982–1991, 2023. 3

    1982

  38. [46]

    Real-time driving risk assessment using deep learning with xgboost

    Liang Shi, Chen Qian, and Feng Guo. Real-time driving risk assessment using deep learning with xgboost. Accident Analysis & Prevention, 178:106836, 2022. 2, 6, 7

    2022

  39. [47]

    Dust: Dual swin transformer for multi-modal video and time-series modeling

    Liang Shi, Yixin Chen, Meimei Liu, and Feng Guo. Dust: Dual swin transformer for multi-modal video and time-series modeling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2024. 6, 7

    2024

  40. [48]

    Scvlm: Enhancing vision-language model for safety-critical event understanding

    Liang Shi, Boyu Jiang, Tong Zeng, and Feng Guo. Scvlm: Enhancing vision-language model for safety-critical event understanding. In Proceedings of the Winter Conference on Applications of Computer Vision, pages 1061–1071, 2025. 7

    2025

  41. [49]

    Unsafe maneuver classification from dashcam video and gps/imu sensors using spatio-temporal attention selector

    Matteo Simoncini, Douglas Coimbra de Andrade, Leonardo Taccari, Samuele Salti, Luca Kubin, Fabio Schoen, and Francesco Sambo. Unsafe maneuver classification from dashcam video and gps/imu sensors using spatio-temporal attention selector. IEEE Transactions on Intelligent Trans-...

    2022

  42. [50]

    Diffusion-based sound synthe- sis in music production

    Pierre-Louis Wolfgang L ´eon Suckrow, Christoph Johannes Weber, and Sylvia Rothe. Diffusion-based sound synthe- sis in music production. In Proceedings of the 12th ACM SIGPLAN International Workshop on Functional Art, Music, Modelling, and Design, pages 55–64, 2024. 3

    2024

  43. [51]

    Scalability in perception for autonomous driving: Waymo open dataset

    Pei Sun, Henrik Kretzschmar, Xerxes Dotiwalla, Aurelien Chouard, Vijaysai Patnaik, Paul Tsui, James Guo, Yin Zhou, Yuning Chai, Benjamin Caine, et al. Scalability in perception for autonomous driving: Waymo open dataset. In Proceed- ings of the IEEE/CVF conference on computer ...

    2020

  44. [52]

    Classification of crash and near-crash events from dashcam videos and telematics

    Leonardo Taccari, Francesco Sambo, Luca Bravi, Samuele Salti, Leonardo Sarti, Matteo Simoncini, and Alessandro Lori. Classification of crash and near-crash events from dashcam videos and telematics. In 2018 21st International Conference on intelligent transportation systems (I...

    2018

  45. [53]

    Moondream2, 2024

    Vikhyat. Moondream2, 2024. Accessed: 2025-02-21. 8

    2024

  46. [54]

    Exploiting diffusion prior for real-world image super-resolution

    Jianyi Wang, Zongsheng Yue, Shangchen Zhou, Kelvin CK Chan, and Chen Change Loy. Exploiting diffusion prior for real-world image super-resolution. International Journal of Computer Vision, 132(12):5929–5949, 2024. 4, 5

    2024

  47. [55]

    Motionsc: Data set and network for real- time semantic mapping in dynamic environments

    Joey Wilson, Jingyu Song, Yuewei Fu, Arthur Zhang, An- drew Capodieci, Paramsothy Jayakumar, Kira Barton, and Maani Ghaffari. Motionsc: Data set and network for real- time semantic mapping in dynamic environments. IEEE Robotics and Automation Letters, 7(3):8439–8446, 2022. 2

    2022

  48. [56]

    Using telematics data to find risky driver be- haviour

    Manda Winlaw, Stefan H Steiner, R Jock MacKay, and Al- laa R Hilal. Using telematics data to find risky driver be- haviour. Accident Analysis & Prevention , 131:131–136,

  49. [57]

    Comfyui-stablesr: Integrating stablesr into com- fyui, 2025

    WSJUSA. Comfyui-stablesr: Integrating stablesr into com- fyui, 2025. Accessed: 2025-02-22. 5

    2025

  50. [58]

    A survey on video dif- fusion models

    Zhen Xing, Qijun Feng, Haoran Chen, Qi Dai, Han Hu, Hang Xu, Zuxuan Wu, and Yu-Gang Jiang. A survey on video dif- fusion models. ACM Computing Surveys, 57(2):1–42, 2024. 3

    2024

  51. [59]

    Fine-grained image editing using controlnet: Expanding pos- sibilities in visual manipulation

    Longfei Xu, Hongbo Huang, Yushuang Zhao, Shuwen Pan, Yaolin Zheng, Xiaoxu Yan, Linkai Huang, and Lishan Wu. Fine-grained image editing using controlnet: Expanding pos- sibilities in visual manipulation. In International Conference on Intelligent Computing, pages 27–38. Springe...

    2024

  52. [60]

    Magicanimate: Temporally consistent human im- age animation using diffusion model

    Zhongcong Xu, Jianfeng Zhang, Jun Hao Liew, Hanshu Yan, Jia-Wei Liu, Chenxu Zhang, Jiashi Feng, and Mike Zheng Shou. Magicanimate: Temporally consistent human im- age animation using diffusion model. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Reco...

    2024

  53. [61]

    Artfu- sion: A diffusion model-based style synthesis framework for portraits

    Hyemin Yang, Heekyung Yang, and Kyungha Min. Artfu- sion: A diffusion model-based style synthesis framework for portraits. Electronics, 13(3):509, 2024. 3

    2024

  54. [62]

    Unsupervised traffic accident detection in first-person videos

    Yu Yao, Mingze Xu, Yuchen Wang, David J Crandall, and Ella M Atkins. Unsupervised traffic accident detection in first-person videos. In 2019 IEEE/RSJ International Confer- ence on Intelligent Robots and Systems (IROS) , pages 273–

    2019

  55. [63]

    Minicpm-v: A gpt-4v level mllm on your phone, 2024

    Yuan Yao, Tianyu Yu, Ao Zhang, Chongyi Wang, Junbo Cui, Hongji Zhu, Tianchi Cai, Haoyu Li, Weilin Zhao, Zhihui He, Qianyu Chen, Huarong Zhou, Zhensheng Zou, Haoye Zhang, Shengding Hu, Zhi Zheng, Jie Zhou, Jie Cai, Xu Han, Guoyang Zeng, Dahai Li, Zhiyuan Liu, and Maosong Sun. M...

    2024

  56. [64]

    Ip- adapter: Text compatible image prompt adapter for text-to- image diffusion models

    Hu Ye, Jun Zhang, Sibo Liu, Xiao Han, and Wei Yang. Ip- adapter: Text compatible image prompt adapter for text-to- image diffusion models. arXiv preprint arXiv:2308.06721,

    arXiv

  57. [65]

    Bdd100k: A diverse driving dataset for heterogeneous multitask learning

    Fisher Yu, Haofeng Chen, Xin Wang, Wenqi Xian, Yingying Chen, Fangchen Liu, Vashisht Madhavan, and Trevor Dar- rell. Bdd100k: A diverse driving dataset for heterogeneous multitask learning. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition ...

    2020

  58. [66]

    Layoutdiffusion: Improving graphic lay- out generation by discrete diffusion probabilistic models

    Junyi Zhang, Jiaqi Guo, Shizhao Sun, Jian-Guang Lou, and Dongmei Zhang. Layoutdiffusion: Improving graphic lay- out generation by discrete diffusion probabilistic models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 7226–7236, 2023. 3

    2023

  59. [67]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023. 3, 4

    2023

  60. [68]

    Weinberger, and Yoav Artzi

    Tianyi Zhang*, Varsha Kishore*, Felix Wu*, Kilian Q. Weinberger, and Yoav Artzi. Bertscore: Evaluating text gen- eration with bert. In International Conference on Learning Representations, 2020. 8

    2020

  61. [69]

    Denoising dif- fusion models for plug-and-play image restoration

    Yuanzhi Zhu, Kai Zhang, Jingyun Liang, Jiezhang Cao, Bi- han Wen, Radu Timofte, and Luc Van Gool. Denoising dif- fusion models for plug-and-play image restoration. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1219–1229, 2023. 3

    2023

  62. [70]

    Wayvescenes101: A dataset and benchmark for novel view synthesis in autonomous driving

    Jannik Z ¨urn, Paul Gladkov, Sof ´ıa Dudas, Fergal Cotter, Sofi Toteva, Jamie Shotton, Vasiliki Simaiaki, and Nikhil Mohan. Wayvescenes101: A dataset and benchmark for novel view synthesis in autonomous driving. arXiv preprint arXiv:2407.08280, 2024. 2 11 Appendix: Variable di...

    2024 arXiv

  63. [2017]

    Technical report, 2017. 1

    2017

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.