Pith. sign in

REVIEW 4 major objections 5 minor 42 references

MapDiffusion: Generative Diffusion for Vectorized Online HD Map Construction and Uncertainty Estimation in Autonomous Driving

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read MapDiffusion's claim: full generative diffusion over vectorized polylines beats the StreamMapNet baseline and turns sample spread into an uncertainty signal.

desk verdict A solid, well-engineered application of diffusion to vectorized HD map construction; the mAP gain is plausible but rests on a non-reproduced baseline, and the uncertainty story is suggestive rather than proven. read the letter →

arxiv 2507.21423 v1 pith:3Y76XXPZ submitted 2025-07-29 cs.CV cs.AIcs.LGcs.RO

Thomas Monninger , Zihan Zhang , Zhipeng Mo , Md Zafar Anwar , Steffen Staab , Sihao Ding This is my paper · ORCID
classification cs.CVcs.AIcs.LGcs.RO
keywords onlineHDmapconstructiondiffusionmodelsvectorizeduncertaintyestimationbird's-eyeviewautonomousdrivingnuScenesDDIMsampling
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 tries to establish that online HD map construction can be formulated as a generative diffusion problem rather than a deterministic prediction problem. MapDiffusion denoises randomly initialized polylines conditioned on a BEV latent grid, so one forward pass yields a sample from the map distribution instead of a single committed map. The experiments report 35.6% mAP versus 33.8% for StreamMapNet, a 5.3% relative gain, at 8.0 FPS with five diffusion steps, and improved ROC AUC from 0.89 to 0.92 when ten samples are aggregated. The same samples give a spatial uncertainty estimate whose mean is 31% higher in occluded regions, suggesting that ambiguity is quantified where it matters. This matters because a self-driving planner can use both the map distribution and its uncertainty rather than trusting one deterministic interpretation.

What carries the argument

The central mechanism is the graph diffusion decoder: a DETR-style transformer decoder turned into a denoiser. It starts from randomly initialized polylines, takes a diffusion time-step embedding, and queries the latent BEV grid through deformable cross-attention; a cosine noise scheduler defines the forward process, and training minimizes a line reconstruction loss plus a class loss. This mechanism is what converts map construction from a single deterministic prediction into sampling from a map distribution, and the same sampled polylines feed both the aggregated prediction and the variance-based uncertainty map.

What would settle it

Rerun StreamMapNet with the identical training recipe and split used for MapDiffusion; if its reproduced mAP reaches or exceeds 35.6%, the claimed 5.3% relative improvement and the state-of-the-art conclusion are unsupported. A second check is to apply the variance-based uncertainty map to scenes with no occlusion; if variance remains high on fully visible geometry, then sample spread is not tracking scene ambiguity.

Watch

Extended reading notes

Core claim

MapDiffusion's central claim is that full generative diffusion over vectorized polylines, conditioned on a BEV latent grid, is a viable replacement for deterministic vector decoding in online HD map construction, and that the resulting sample distribution carries a usable uncertainty signal. On the nuScenes split without geospatial overlap, the single-sample prediction reaches 35.6% mAP against 33.8% for StreamMapNet, a 5.3% relative gain, at 8.0 FPS with five DDIM steps. Sampling ten maps and aggregating them in raster space improves ROC AUC from 0.89 to 0.92. The per-location variance across samples is the uncertainty estimate, and it is 31% higher in invisible than in visible areas, a difference the paper reports as significant at p < 0.001.

Load-bearing premise

The 33.8% mAP StreamMapNet baseline is taken from the original paper and was not reproduced under identical training conditions; the claimed 5.3% improvement depends on that number being directly comparable to the MapDiffusion run, and if the baseline were rerun and scored higher, the headline gain would shrink or disappear.

Editorial extensions

If this is right

  • Aggregating 10 rasterized samples raises ROC AUC from 0.89 to 0.92 relative to a single sample, so sampling is an accuracy lever, not just a visualization tool.
  • The variance across samples is significantly higher in invisible areas, with mean 0.0082 versus 0.0063, giving a spatial uncertainty signal that tracks occlusion.
  • MapDiffusion reaches 35.6% mAP without learned queries or decoder-side temporal aggregation, so the generative decoder can replace the query mechanism rather than merely refine it.
  • Because the BEV encoder runs once, inference stays real-time: 8.0 FPS with five diffusion steps and 12.8 FPS with one step on the reported hardware.
  • Class prediction is excluded from the diffusion loop, so denoising happens purely in vector space, which is what lets the model generate vectorized maps directly instead of raster masks.

Reading between the lines

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

  • The paper does not evaluate downstream planning, but feeding the per-location variance into a trajectory predictor or planner is a direct testable extension; higher variance in occluded areas should reduce confidence in any single map-based plan.
  • Because the BEV encoder is computed once and only the decoder is iterated, the same diffusion decoder could be attached to newer non-diffusion map heads; if so, those models would inherit sample diversity and uncertainty without redesigning their encoder.
  • The 31% gap between visible and invisible uncertainty likely understates the effect of occlusion, since the StreamMapNet-style BEV temporal aggregation leaks information from previous frames into currently occluded areas; a single-frame variant should show a larger gap.
  • A useful stress test is to check whether sample variance in occluded regions reflects genuine topological ambiguity, such as one lane versus two, rather than just noise; scoring samples by distinct topologies would separate those cases.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. MapDiffusion proposes to replace the deterministic DETR-style query decoder of StreamMapNet with a diffusion-based decoder that iteratively denoises randomly initialized polylines conditioned on a BEV latent grid. The model is trained with a noise scheduler on vectorized map elements and sampled with DDIM, producing multiple map hypotheses. The paper reports 35.6% mAP on the nuScenes split (vs. an asterisked 33.8% for StreamMapNet), shows that aggregating 10 rasterized samples improves ROC AUC from 0.89 to 0.92, and computes a spatial uncertainty map U from cross-sample variance, which is 31% higher in invisible than in visible areas (p<0.001). Ablations cover diffusion steps, DDIM eta, query threshold, BEV encoder pretraining, and query padding strategies.

Significance. If the accuracy and uncertainty claims hold, MapDiffusion is a useful demonstration that a full generative diffusion process can operate directly on vectorized map elements for online HD map construction, rather than on raster maps or as a refinement of existing proposals. The paper is explicit in comparing against PolyDiffuse and positions the uncertainty estimate as the main novelty. The ablations are thorough regarding diffusion parameters and padding strategies, and the sample-aggregation experiment provides a concrete check that multiple samples carry information. However, the current evidence for the headline accuracy gain relies on a non-reproduced baseline, and the uncertainty validation only shows a correlation with a visibility proxy rather than with map error, leaving the decision-relevant claim under-supported. The paper's strengths are its clear formulation of the diffusion setup over polylines and the recognition that sample variance is a natural uncertainty signal.

major comments (4)
  1. [Section IV-D, Table I] The central accuracy claim (35.6% vs. 33.8% mAP, a 5.3% relative improvement) is computed against an asterisked StreamMapNet baseline taken from the original paper [8], not reproduced with the same training pipeline, evaluation code, or hardware. Because MapDiffusion changes the decoder, query initialization, and training procedure, the comparability of the two numbers is not established. Please reproduce StreamMapNet under identical conditions and, for both models, report mean and standard deviation over multiple seeds; the mAP differences in Table II (35.2 to 35.6) are small enough that single-run variability could affect the conclusion.
  2. [Section IV-E.2, Eq. (4)] The uncertainty evaluation demonstrates only that U is higher in invisible/occluded areas (mean 0.0082 vs. 0.0063, 31% higher, p<0.001). It does not test whether U correlates with map prediction error, which is the decision-relevant quantity claimed in the abstract and conclusion. A stochastic model whose output noise is larger in occluded regions would pass this test without providing useful uncertainty for planning. Please add a direct calibration analysis (for example, bin U and measure per-pixel or per-instance error against GT maps), compare against Monte Carlo dropout or random query perturbation, and preferably evaluate a downstream use as in Gu et al. [30]. Without such evidence, the phrase 'uncertainty-aware decision-making' in the conclusion is not supported.
  3. [Section IV-E.1, Section III-C.1] The aggregation benefit is demonstrated only in raster space, using the ROC/AUC of rasterized probability maps (Eqs. 1-3). The output representation used by planning systems is the vectorized map, and the paper does not show that aggregating samples improves vectorized instance-level metrics such as mAP or the quality of the final polyline set. Since Section III-C.1 states that raster-space aggregation is 'sufficient to demonstrate our point', please make explicit that the claimed accuracy improvement from aggregation is restricted to rasterized evaluation, or provide vectorized aggregation results.
  4. [Section IV-G.1, Table II] The model reaches 35.2% mAP with a single diffusion step and 35.6% with five steps, so the accuracy gain from the iterative diffusion process itself is small. Since the StreamMapNet baseline also uses iterative refinement in its decoder, the specific contribution of the diffusion mechanism over a deterministic refinement scheme is not isolated. Please add an ablation with the same decoder but deterministic refinement or with learned queries, and report sample-diversity statistics (e.g., pairwise distance between sampled maps) to substantiate that the model is generating a meaningful distribution rather than noise around a single mode.
minor comments (5)
  1. [Section IV-D] The sentence 'Table I shows the qualitative results' should read 'quantitative results'.
  2. [Section IV-E.2] Please clarify the unit of analysis for the t-test: the text says 'mean uncertainty per traffic scene' but the figure shows a distribution of variances; if the test is per-pixel, spatial autocorrelation would make the p-value invalid.
  3. [Section IV-G.3] The 'Smooth' padding strategy is described only as 'smooth random polylines (both straight and curved) or polygons'; the generation procedure should be specified or referenced.
  4. [Section IV-B] Inference FPS is reported on an NVIDIA A10 while training is on V100s; please state the GPU used for each FPS number and whether timing includes the BEV encoder.
  5. [Section III-C.2, Eq. (4)] The spatial resolution H x W of the uncertainty map U is not specified; please state whether it matches the BEV grid (100x50) or another resolution.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: MapDiffusion is benchmarked against external nuScenes ground truth and its uncertainty maps are validated against independent Occ3D-derived visibility masks; the only flagged issue (non-reproduced StreamMapNet baseline) is a comparability risk, not a circular reduction.

full rationale

The paper's central accuracy claim (35.6% mAP vs. 33.8% StreamMapNet baseline, Table I) is an empirical comparison against externally provided nuScenes vectorized annotations, and the MapDiffusion model is trained with a standard diffusion objective (minimizing line loss between the denoised prediction p_theta(q(x0,t),t,B) and the ground-truth x0, Section III-B.2). No parameter is fitted to the test set and then renamed as a prediction; the n-sample aggregation is evaluated against the same external GT via ROC/AUC (Section IV-E.1). The uncertainty estimate U is defined as sample variance of rasterized class probabilities (Eqs. 1-4), and the paper's validation is an external correlation test: mean U in Occ3D-ray-traced invisible regions is significantly higher than in visible regions (Section IV-E.2, Fig. 5). This is an empirical finding, not an identity; visibility masks and GT maps are not inputs to U. The only self-citations ([6] TempBEV, [13] AugMapNet) appear in related-work descriptions and are not load-bearing for the derivation. The asterisked StreamMapNet baseline in Table I is taken from the original paper rather than reproduced, so the 5.3% improvement could shift under identical retraining; that is a correctness/comparability concern, not a circularity, because the comparison is still to an external model and dataset. Overall, the derivation chain is self-contained and externally checkable, so the circularity score is low.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The empirical claims rest on assumptions about the sufficiency of the BEV condition, the validity of diffusion on vector coordinates, the accuracy of proxy visibility masks, and raster-space aggregation as a proxy for vector-space fusion. No new theoretical entities are introduced.

free parameters (7)
  • Number of diffusion steps k = 5
    Chosen by ablation (Table II); performance saturates at k=5.
  • DDIM eta = 0.5
    Chosen by ablation (Table II); minor influence on performance.
  • Query threshold tau = 0.5
    Chosen by ablation (Table II); strongest hyperparameter influence on mAP.
  • Padding Gaussian mean/std = mu=0.5, sigma=0.25
    Chosen by design and padding ablation (Table III); clipped to [0,1].
  • Gaussian filter size/scale for aggregation = g=5 (3m), sigma=1 (0.6m)
    Arbitrary smoothing choices in sample aggregation, Section IV-B.
  • Rasterize score threshold = 0.4
    Score threshold for rasterizing polylines in aggregation, Section IV-B.
  • Number of samples n = 10
    Set for multi-sample experiments; no ablation on n.
assumptions (5)
  • domain assumption The BEV latent grid, computed from camera images, contains sufficient spatial information to condition the diffusion decoder.
    The entire method conditions denoising on B; the paper provides no analysis of how much condition information is present after the BEV encoder or whether failure cases are due to missing condition. See Section III-B.3.
  • domain assumption Denoising Diffusion Probabilistic Models, applied to polyline coordinates with a line loss, learn the intended distribution of vectorized maps.
    The paper assumes DDPM/DDIM theory transfers to structured vector outputs without analyzing the learned distribution; no evidence of distribution quality beyond qualitative samples. See Section III-B.2.
  • domain assumption Visibility masks generated from Occ3D occupancy maps accurately represent which map locations are occluded from the ego cameras.
    The uncertainty validation compares sample variance against these masks; ground and flat voxels are excluded and ray tracing is applied to a 2D projection, an indirect procedure. See Section IV-E.2.
  • domain assumption Aggregation in raster space is a sufficient proxy for evaluating the benefit of multiple vectorized map samples.
    Authors state raster space is 'sufficient to demonstrate our point' (Section IV-E.1), but downstream planning requires vectorized maps; the vector-level benefit is not measured.
  • domain assumption The nuScenes split without geospatial overlap from StreamMapNet is a fair evaluation protocol and all baselines are evaluated under comparable conditions.
    The paper uses this split but only reproduces SQD-MapNet; StreamMapNet number is taken from the original paper. The fairness of the comparison is load-bearing for the SOTA claim. See Section IV-D.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MapDiffusion: Generative Diffusion for Vectorized Online HD Map Construction and Uncertainty Estimation in Autonomous Driving." pith.science (2026). https://pith.science/paper/3Y76XXPZ

@misc{pith2026250721423,
  author       = {Pith},
  title        = {Pith review of: MapDiffusion: Generative Diffusion for Vectorized Online HD Map Construction and Uncertainty Estimation in Autonomous Driving},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3Y76XXPZ}},
  note         = {Machine review of arXiv:2507.21423}
}
read the original abstract

Autonomous driving requires an understanding of the static environment from sensor data. Learned Bird's-Eye View (BEV) encoders are commonly used to fuse multiple inputs, and a vector decoder predicts a vectorized map representation from the latent BEV grid. However, traditional map construction models provide deterministic point estimates, failing to capture uncertainty and the inherent ambiguities of real-world environments, such as occlusions and missing lane markings. We propose MapDiffusion, a novel generative approach that leverages the diffusion paradigm to learn the full distribution of possible vectorized maps. Instead of predicting a single deterministic output from learned queries, MapDiffusion iteratively refines randomly initialized queries, conditioned on a BEV latent grid, to generate multiple plausible map samples. This allows aggregating samples to improve prediction accuracy and deriving uncertainty estimates that directly correlate with scene ambiguity. Extensive experiments on the nuScenes dataset demonstrate that MapDiffusion achieves state-of-the-art performance in online map construction, surpassing the baseline by 5% in single-sample performance. We further show that aggregating multiple samples consistently improves performance along the ROC curve, validating the benefit of distribution modeling. Additionally, our uncertainty estimates are significantly higher in occluded areas, reinforcing their value in identifying regions with ambiguous sensor input. By modeling the full map distribution, MapDiffusion enhances the robustness and reliability of online vectorized HD map construction, enabling uncertainty-aware decision-making for autonomous vehicles in complex environments.

Figures

Figures reproduced from arXiv: 2507.21423 by the authors.

Figure 1
Figure 1. Overview of the diffusion forward and backward [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of MapDiffusion approach based on schematic traffic scene with an occluded camera view on the ego’s [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of StreamMapNet, MapDiffusion during inference, and MapDiffusion during training, with box colors [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: ROC curves for one sample (red) and 10 aggregated [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Relation between visibility and variance across pre [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Two qualitative results of MapDiffusion. The top two rows show the 6 camera views. The third row shows the GT [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 39 canonical work pages

  1. [8]

    Streammapnet: Streaming mapping network for vectorized online hd map construc- tion,

    T. Yuan, Y . Liu, Y . Wang, Y . Wang, and H. Zhao, “Streammapnet: Streaming mapping network for vectorized online hd map construc- tion,” in Proceedings of the IEEE/CVF Winter Conference on Appli- cations of Computer Vision , 2024, pp. 7356–7365. 1, 2, 3, 5

    work page 2024

  2. [30]

    Pro- ducing and leveraging online map uncertainty in trajectory prediction,

    X. Gu, G. Song, I. Gilitschenski, M. Pavone, and B. Ivanovic, “Pro- ducing and leveraging online map uncertainty in trajectory prediction,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 14 521–14 530. 2

    work page 2024

  3. [1]

    Vectormapnet: End-to-end vectorized hd map learning,

    Y . Liu, T. Yuan, Y . Wang, Y . Wang, and H. Zhao, “Vectormapnet: End-to-end vectorized hd map learning,” in International Conference on Machine Learning . PMLR, 2023, pp. 22 352–22 369. 1, 2, 5

    work page 2023

  4. [2]

    Maptrv2: An end-to-end framework for online vectorized hd map construction,

    B. Liao, S. Chen, Y . Zhang, B. Jiang, Q. Zhang, W. Liu, C. Huang, and X. Wang, “Maptrv2: An end-to-end framework for online vectorized hd map construction,” International Journal of Computer Vision, 2025. 1, 5

    work page 2025

  5. [3]

    Vectornet: Encoding hd maps and agent dynamics from vectorized representation,

    J. Gao, C. Sun, H. Zhao, Y . Shen, D. Anguelov, C. Li, and C. Schmid, “Vectornet: Encoding hd maps and agent dynamics from vectorized representation,” in Proceedings of the IEEE/CVF conference on com- puter vision and pattern recognition , 2020, pp. 11 525–11 533. 1

    work page 2020

  6. [4]

    Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,

    J. Philion and S. Fidler, “Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XIV 16. Springer, 2020, pp. 194–210. 1, 2

    work page 2020

  7. [5]

    Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers,

    Z. Li, W. Wang, H. Li, E. Xie, C. Sima, T. Lu, Y . Qiao, and J. Dai, “Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers,” in ECCV. Springer, 2022, pp. 1–18. 1, 2

    work page 2022

  8. [6]

    TempBEV: Improving learned bev encoders with combined image and bev space temporal aggregation,

    T. Monninger, V . Dokkadi, M. Z. Anwar, and S. Staab, “TempBEV: Improving learned bev encoders with combined image and bev space temporal aggregation,” in 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems . IEEE, 2024, pp. 9668–9675. 1, 2

    work page 2024

Show all 42 references
  1. [7]

    Petrv2: A unified framework for 3d perception from multi-camera images,

    Y . Liu, J. Yan, F. Jia, S. Li, A. Gao, T. Wang, and X. Zhang, “Petrv2: A unified framework for 3d perception from multi-camera images,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 3262–3272. 1, 2

    2023

  2. [9]

    Diffmap: Enhancing map segmentation with map prior using diffusion model,

    P. Jia, T. Wen, Z. Luo, M. Yang, K. Jiang, Z. Liu, X. Tang, Z. Lei, L. Cui, B. Zhang, K. Sheng, and D. Yang, “Diffmap: Enhancing map segmentation with map prior using diffusion model,” IEEE Robotics and Automation Letters , vol. 9, no. 11, pp. 9836–9843, 2024. 1, 2

    2024

  3. [10]

    Diffusion model for ro- bust multi-sensor fusion in 3d object detection and bev segmentation,

    D.-T. Le, H. Shi, J. Cai, and H. Rezatofighi, “Diffusion model for ro- bust multi-sensor fusion in 3d object detection and bev segmentation,” in ECCV. Springer, 2024, pp. 232–249. 1, 2

    2024

  4. [11]

    Polydiffuse: Polygonal shape reconstruction via guided set diffusion models,

    J. Chen, R. Deng, and Y . Furukawa, “Polydiffuse: Polygonal shape reconstruction via guided set diffusion models,” Advances in neural information processing systems , vol. 36, 2023. 1, 2, 5

    2023

  5. [12]

    Maptr: Structured modeling and learning for online vectorized hd map construction,

    B. Liao, S. Chen, X. Wang, T. Cheng, Q. Zhang, W. Liu, and C. Huang, “Maptr: Structured modeling and learning for online vectorized hd map construction,” in International Conference on Learning Repre- sentations, 2023. 2, 5

    2023

  6. [13]

    Augmap- net: Improving spatial latent structure via bev grid augmentation for enhanced vectorized online hd map construction,

    T. Monninger, M. Z. Anwar, S. Antol, S. Staab, and S. Ding, “Augmap- net: Improving spatial latent structure via bev grid augmentation for enhanced vectorized online hd map construction,” arXiv preprint arXiv:2503.13430, 2025. 2

    2025

  7. [14]

    Stream query denoising for vectorized hd- map construction,

    S. Wang, F. Jia, W. Mao, Y . Liu, Y . Zhao, Z. Chen, T. Wang, C. Zhang, X. Zhang, and F. Zhao, “Stream query denoising for vectorized hd- map construction,” in ECCV. Springer, 2024, pp. 203–220. 2, 5

    2024

  8. [15]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems , vol. 33,

  9. [16]

    Generative modeling by estimating gradients of the data distribution,

    Y . Song and S. Ermon, “Generative modeling by estimating gradients of the data distribution,” Advances in neural information processing systems, vol. 32, 2019. 2

    2019

  10. [17]

    Diffusion models beat gans on image synthesis,

    P. Dhariwal and A. Nichol, “Diffusion models beat gans on image synthesis,” Advances in neural information processing systems, vol. 34,

  11. [18]

    DiffusionDet: Diffusion model for object detection,

    S. Chen, P. Sun, Y . Song, and P. Luo, “DiffusionDet: Diffusion model for object detection,” in ICCV, October 2023, pp. 19 830–19 843. 2

    2023

  12. [19]

    Score-based generative modeling through stochastic differential equations,

    Y . Song, J. Sohl-Dickstein, D. P. Kingma, A. Kumar, S. Ermon, and B. Poole, “Score-based generative modeling through stochastic differential equations,” in International Conference on Learning Rep- resentations, 2021. 2

    2021

  13. [20]

    Hierarchi- cal text-conditional image generation with clip latents,

    A. Ramesh, P. Dhariwal, A. Nichol, C. Chu, and M. Chen, “Hierarchi- cal text-conditional image generation with clip latents,” arXiv preprint arXiv:2204.06125, 2022. 2

    2022 arXiv

  14. [21]

    Occworld: Learning a 3d occupancy world model for autonomous driving,

    W. Zheng, W. Chen, Y . Huang, B. Zhang, Y . Duan, and J. Lu, “Occworld: Learning a 3d occupancy world model for autonomous driving,” in ECCV. Springer, 2024, pp. 55–72. 2

    2024

  15. [22]

    Occgen: Generative multi-modal 3d occupancy prediction for autonomous driving,

    G. Wang, Z. Wang, P. Tang, J. Zheng, X. Ren, B. Feng, and C. Ma, “Occgen: Generative multi-modal 3d occupancy prediction for autonomous driving,” in ECCV. Springer, 2024, pp. 95–112. 2

    2024

  16. [23]

    Uniscene: Unified occupancy-centric driving scene generation,

    B. Li, J. Guo, H. Liu, Y . Zou, Y . Ding, X. Chen, H. Zhu, F. Tan, C. Zhang, T. Wang et al. , “Uniscene: Unified occupancy-centric driving scene generation,” arXiv preprint arXiv:2412.05435 , 2024. 2

    2024 arXiv

  17. [24]

    Online diffusion-based 3d occupancy prediction at the frontier with proba- bilistic map reconciliation,

    A. Reed, L. Achey, B. Crowe, B. Hayes, and C. Heckman, “Online diffusion-based 3d occupancy prediction at the frontier with proba- bilistic map reconciliation,” arXiv preprint arXiv:2409.10681 , 2024. 2

    2024 arXiv

  18. [25]

    Generating urban road networks with conditional diffusion models,

    X. Gu, M. Zhang, J. Lyu, and Q. Ge, “Generating urban road networks with conditional diffusion models,” ISPRS International Journal of Geo-Information, vol. 13, no. 6, p. 203, 2024. 2

    2024

  19. [26]

    Lane segmentation refinement with diffusion models,

    A. Ruiz, A. Melnik, D. Wang, and H. Ritter, “Lane segmentation refinement with diffusion models,” arXiv preprint arXiv:2405.00620 ,

  20. [27]

    Radiodiff: An effective generative diffusion model for sampling-free dynamic radio map construction,

    X. Wang, K. Tao, N. Cheng, Z. Yin, Z. Li, Y . Zhang, and X. Shen, “Radiodiff: An effective generative diffusion model for sampling-free dynamic radio map construction,” IEEE Transactions on Cognitive Communications and Networking , 2024. 2

    2024

  21. [28]

    Diffbev: Conditional diffusion model for bird’s eye view perception,

    J. Zou, K. Tian, Z. Zhu, Y . Ye, and X. Wang, “Diffbev: Conditional diffusion model for bird’s eye view perception,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 38, no. 7, 2024, pp. 7846–7854. 2

    2024

  22. [29]

    Housediffusion: Vector floorplan generation via a diffusion model with discrete and continuous denoising,

    M. A. Shabani, S. Hosseini, and Y . Furukawa, “Housediffusion: Vector floorplan generation via a diffusion model with discrete and continuous denoising,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 5466–5475. 2

    2023

  23. [31]

    Card: Classification and regression diffusion models,

    X. Han, H. Zheng, and M. Zhou, “Card: Classification and regression diffusion models,” in Thirty-Sixth Conference on Neural Information Processing Systems, 2022. 2

    2022

  24. [32]

    Diffusion-based probabilistic uncertainty estimation for active domain adaptation,

    Z. Du and J. Li, “Diffusion-based probabilistic uncertainty estimation for active domain adaptation,” Advances in Neural Information Pro- cessing Systems, vol. 36, pp. 17 129–17 155, 2023. 2

    2023

  25. [33]

    Motiondiffuser: Controllable multi-agent motion prediction using diffusion,

    C. Jiang, A. Cornman, C. Park, B. Sapp, Y . Zhou, D. Anguelov et al. , “Motiondiffuser: Controllable multi-agent motion prediction using diffusion,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 9644–9653. 2

    2023

  26. [34]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in ECCV. Springer, 2020, pp. 213–229. 3

    2020

  27. [35]

    Deformable DETR: Deformable transformers for end-to-end object detection,

    X. Zhu, W. Su, L. Lu, B. Li, X. Wang, and J. Dai, “Deformable DETR: Deformable transformers for end-to-end object detection,” in International Conference on Learning Representations , 2021. 3

    2021

  28. [36]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” in International Conference on Learning Representations ,

  29. [37]

    nuScenes: A multimodal dataset for autonomous driving,

    H. Caesar, V . Bankiti, A. H. Lang, S. V ora, V . E. Liong, Q. Xu, A. Krishnan, Y . Pan, G. Baldan, and O. Beijbom, “nuScenes: A multimodal dataset for autonomous driving,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020. 5

    2020

  30. [38]

    Online map vectorization for autonomous driving: A rasterization perspective,

    G. Zhang, J. Lin, S. Wu, Z. Luo, Y . Xue, S. Lu, Z. Wang et al. , “Online map vectorization for autonomous driving: A rasterization perspective,” Advances in Neural Information Processing Systems , vol. 36, 2024. 5

    2024

  31. [39]

    Mgmap: Mask-guided learning for online vectorized hd map construction,

    X. Liu, S. Wang, W. Li, R. Yang, J. Chen, and J. Zhu, “Mgmap: Mask-guided learning for online vectorized hd map construction,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 14 812–14 821. 5

    2024

  32. [40]

    Leveraging enhanced queries of point sets for vectorized map construction,

    Z. Liu, X. Zhang, G. Liu, J. Zhao, and N. Xu, “Leveraging enhanced queries of point sets for vectorized map construction,” in ECCV, 2024. 5

    2024

  33. [41]

    Himap: Hybrid representation learning for end-to-end vectorized hd map construction,

    Y . Zhou, H. Zhang, J. Yu, Y . Yang, S. Jung, S.-I. Park, and B. Yoo, “Himap: Hybrid representation learning for end-to-end vectorized hd map construction,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024. 5

    2024

  34. [42]

    Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving,

    X. Tian, T. Jiang, L. Yun, Y . Wang, Y . Wang, and H. Zhao, “Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving,” Advances in neural information processing systems , vol. 36,

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.