Pith. sign in

REVIEW 5 major objections 4 minor 1 cited by

SparseMeXT Unlocking the Potential of Sparse Representations for HD Map Construction

T0 review · 5 major / 4 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 SparseMeXt is claimed to be the first sparse HD-map construction model to surpass dense BEV methods, reaching 68.9% mAP on nuScenes at over 20 FPS.

desk verdict Useful sparse-HD-map engineering with three new components, but the headline claim of being the first sparse method to beat dense approaches is unsupported by the paper's own matched-backbone numbers. read the letter →

arxiv 2505.08808 v1 pith:4YG2EIHI submitted 2025-05-12 cs.CV cs.AI

Anqing Jiang , Jinhao Chai , Yu Gao , Yiru Wang , Yuwen Heng , Zhigang Sun , Hao Sun , Zezhong Zhao
show 5 more authors
Li Sun Jian Zhou Lijuan Zhu Shugong Xu Hao Zhao
This is my paper · ORCID
classification cs.CVcs.AI
keywords sparserepresentationsHDmapconstructionnuScenesquerydenoisingauxiliarysegmentationvectorizedmapsautonomousdrivingsingle-in-multi-outneck
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

Online high-definition map construction has been dominated by dense bird's-eye-view (BEV) representations that are computationally heavy. This paper argues that sparse, query-based representations can close that gap and exceed it, provided the architecture is tailored to map geometry instead of borrowed from 3D object detection. It reports SparseMeXt-Large at 68.9% mAP on nuScenes with over 20 FPS, surpassing the dense MapTRv2-Base at 68.7% and the sparse SparseDrive-B at 56.2%. The paper's claim is that sparse representations are not inherently weaker, and its three design changes are what unlock their potential.

What carries the argument

The load-bearing mechanism is a sparse encoder-decoder with map-specific modifications at every level. The image encoder uses a nuImages-pretrained ResNet backbone and a single-in-multi-out (SiMo) neck that feeds only the backbone's C5 feature into the decoder, justified by the observation that map elements occupy consistently large image areas and benefit less from multi-scale feature fusion. The decoder splits into independent classification and regression branches at the deformable feature aggregation layer so each task samples its own points. The auxiliary sparse-dense segmentation head converts instance features into a BEV segmentation map during training only, supplying global foreground supervision without dense BEV computation at inference. Finally, physical-prior query denoising (PPDN) injects rotation, location, scale, and curvature perturbations that respect the geometry of polylines and polygons, stabilizing DETR-style query training for point-sequence outputs.

What would settle it

Rerun the ablation sequence in Table X with at least five random seeds under the same training protocol; if the seed-to-seed spread on any component equals or exceeds the reported gain (0.07% stage ratio, 0.6% decouple-DFA, 0.9% segmentation loss, 0.5% PPDN), the stepwise attribution collapses. Alternatively, an independent reproduction of SparseMeXt-Large on the nuScenes validation split could check whether 68.9% mAP actually exceeds MapTRv2-Base's 68.7% beyond chance.

Watch

Extended reading notes

Core claim

The central discovery is that sparse HD map construction can match and surpass dense BEV methods when each design stage is revisited for map-specific properties. SparseMeXt achieves this through coordinated changes: a backbone pre-trained on driving-domain detection rather than ImageNet classification, replacing the multi-in-multi-out FPN neck with a single-in-multi-out neck because map elements are large objects, adjusting the balance between non-temporal and temporal decoder stages, and decoupling classification from regression in the deformable feature aggregation. On top of the architecture, the paper adds a query-centric sparse-dense auxiliary segmentation loss for global foreground supervision, which costs nothing at inference, and a physical-prior query denoising scheme with rotation, location, scale, and curvature noise tailored to line-shaped map elements. With these components, SparseMeXt-Base reaches 65.2% mAP and SparseMeXt-Large reaches 68.9% mAP, the latter exceeding the dense MapTRv2-Base by 0.2 points while running at over 20 FPS on an RTX 3090.

Load-bearing premise

The load-bearing premise is that the ablation deltas of 0.07 to 0.9 mAP points reported in Section IV-C are real improvements rather than single-run noise; the paper does not report variance or multiple seeds for those steps.

Editorial extensions

If this is right

  • If the 68.9% mAP result holds, sparse map construction is no longer a compromise: it beats the dense state-of-the-art MapTRv2-Base while avoiding dense BEV features, at over 20 FPS.
  • The design principle generalizes: because map elements are large and static, a single-scale neck, a 1+4 stage ratio, and decoupled sampling can be chosen by that geometry rather than by 3D detection defaults.
  • The auxiliary segmentation head and PPDN add accuracy at training time only, so the deployed model keeps the sparse pipeline's speed and memory advantage.
  • Centerline prediction extends the same architecture to 58.8% mAP on nuScenes, connecting map construction to downstream motion prediction and planning.
  • Long-range perception to 60 m by 90 m is feasible with a sparse model, where SparseMeXt-Base outscores MapTR by 7.4% mAP.

Reading between the lines

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

  • Beyond the paper: the SiMo neck result suggests that large-object perception tasks may not need multi-scale fusion at all; lane detection and drivable-area segmentation are natural test beds.
  • Beyond the paper: the physical-prior noise modes (rotation, shift, scale, curvature) are geometry-agnostic and could be applied to trajectory prediction or any polyline regression, where box-style noise is a poor fit.
  • Beyond the paper: the +1.5% mAP attributed to nuImages pre-training is the most portable single finding; a multi-seed study across backbones would tell whether task-relevant pre-training, not architecture, drives the gain.
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

5 major / 4 minor

Summary. The manuscript proposes SparseMeXt, an online HD map construction architecture that retains a sparse, query-based paradigm instead of a dense BEV feature grid. It introduces three main components: an architecture optimized for map features (SiMo image encoder neck, adjusted temporal/non-temporal decoder stage ratio, and a decoupled deformable feature aggregation decoder), a query-centric sparse-dense auxiliary segmentation loss, and a physical-prior query denoising module (PPDN) with rotation, location, scale, and curvature noise. Experiments on nuScenes report SparseMeXt-Tiny/B/L at 55.5%, 65.2%, and 68.9% mAP with 32.9/25.4/20.2 FPS, respectively, plus long-range and centerline detection extensions. The paper's central claim is that SparseMeXt is the first sparse paradigm to surpass dense methods in online HD map construction.

Significance. If the results hold, the paper is a useful demonstration that sparse map construction can be competitive with dense BEV pipelines while remaining efficient, and the component-level analysis identifies concrete design choices worth adopting. The paper's strengths include a structured ablation path from the SparseDrive baseline, same-hardware FPS measurements for several variants, and additional long-range and centerline experiments. The main limitation is that the headline 'surpass dense methods' claim is not currently supported by a controlled comparison at equal backbone capacity, and several component gains are reported from single runs without variance information. These issues are fixable, but they affect the paper's central quantitative narrative.

major comments (5)
  1. [Section IV-B, Table VII] There is an internal inconsistency in the MapTRv2-Base number: Section IV-B states that SparseMeXt-Base outperforms MapTRv2-Base at 63.0%, while Table VII reports MapTRv2-Base at 68.7%. If Table VII is correct, then SparseMeXt-Base (65.2 mAP) is 3.5 points behind the matched-backbone dense method. This inconsistency directly affects the central claim and must be corrected in both the text and the table.
  2. [Section I, Section IV-B, Table VII] The claim that SparseMeXt is the first sparse paradigm to surpass dense methods is not established by the reported comparisons. At the Base/Large scale, the only sparse-over-dense win is SparseMeXt-Large (ResNet-101, 68.9) against MapTRv2-Base (ResNet-50, 68.7), a comparison that is unmatched in backbone capacity; at matched ResNet-50, the dense method is ahead (68.7 vs. 65.2). SparseMeXt-Tiny (ResNet-18, 55.5) does beat MapTRv2-Tiny (ResNet-18, 52.3), so the claim is not vacuous, but a capacity-controlled comparison is needed. Please report a dense baseline trained under the same recipe at matched capacity, or restrict the headline claim to the configuration for which a controlled comparison is available.
  3. [Tables III, IV, V, VI, X; Section IV-C] Several component contributions are supported only by single-run differences of 0.07, 0.6, 0.5, and 0.9 mAP points, with no standard deviation, number of seeds, or significance measure. On nuScenes mAP, differences of this size are commonly within run-to-run noise; as written, the stepwise attribution in Table X and the stage-ratio conclusion in Table III are not firmly supported. Please report at least three seeds with mean and standard deviation for the main ablation path, and re-evaluate whether the 0.07-point stage-ratio gain and the 0.6-point Decouple-DFA gain are reliable.
  4. [Sections I and IV-C; Tables VII and X] The text repeatedly claims a 10.1% mAP improvement over 'SparseDrive', but the numbers do not line up: Table X's 'SparseDrive-Map' baseline is 57.4, Table VII lists SparseDrive-S at 55.1 and SparseDrive-B at 56.2, and SparseMeXt-Base is 65.2. Depending on the baseline, the absolute gain is 7.8, 9.0, or 10.1 mAP points. Please specify the exact baseline configuration and state clearly whether the reported value is an absolute mAP-point gain or a relative percentage gain.
  5. [Section III-D] PPDN is one of the three central contributions, but Section III-D does not provide enough detail to reproduce the denoising scheme: the rotation, scaling, and curvature noise operations are described verbally, without explicit equations, coordinate conventions, noise magnitude ranges, or sampling probabilities. Please add a precise algorithm or equations for the four noise types and report the hyperparameters used in training.
minor comments (4)
  1. [Table III] The row labeled '1 non-temporal, 4 temporal' displays the same rounded average mAP (57.4) as the baseline but is annotated as '+0.07'; either unrounded values are needed or the annotation should be reconciled with the displayed numbers, and the unit should be mAP points rather than percent.
  2. [Section IV-B / Table IX] The text says the long-range setting is '60x90m' and '90m perception range', while Table IX labels rows as '90×60m'. Please make the x/y range convention consistent.
  3. [Throughout] There are numerous typos and inconsistent spellings, including 'TabelX', 'SparseMaXt', 'pdedstrain crossing', 'Nuscenes' vs. 'nuScenes', and 'stage radio' for 'stage ratio'. A careful proofreading pass is needed.
  4. [Section II-A] Reference [6] is cited as 'Persformer: A transformer architecture for topological machine learning', but the same name is used in [18] for a 3D lane detection paper; please verify that the correct references are attached to the textual mentions.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's improvements are empirical ablations against external baselines, not derivations from the claimed result.

full rationale

SparseMeXt is an empirical systems paper. Each claimed component improvement is measured by adding the component to a SparseDrive-based baseline and reporting mAP on the nuScenes validation set; the baseline is an external published method (SparseDrive, [17]), and none of the components is defined in terms of the target mAP. The ablation chain in Table X starts from 'SparseDrive-Map Baseline' and sequentially reports deltas, so the final number is not obtained by fitting a parameter to the benchmark and then predicting the same benchmark. The architectural choices (SiMo from YOLOF, query denoising from DN-DETR, nuImages pretraining) are borrowed from external work or standard practice, and the paper does not invoke a self-citation chain or a uniqueness theorem to force its design. The paper's headline claim that a sparse method surpasses dense methods is supported by one unmatched comparison (SparseMeXt-Large with ResNet-101 versus MapTRv2-Base with ResNet-50), and several ablations report single-run differences as small as 0.07 mAP without variance; those are correctness and rigor concerns, not circularity. No equation or claim in the paper reduces by construction to its own input, so the circularity burden is not met.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claim depends on the nuScenes benchmark, the choice of pretraining, and the detailed training protocol inherited from SparseDrive. The free parameters listed are the main hand-set hyperparameters; the PPDN noise magnitudes are not disclosed, which is a gap in reproducibility. No invented entities are introduced beyond the training-time noise types.

free parameters (8)
  • initial learning rate = 0.006
    AdamW with half-cosine decay; chosen by the authors, not derived.
  • weight decay = 0.01
    Standard choice for AdamW.
  • gradient clipping norm = 35
    Chosen to stabilize training.
  • number of map queries = 100
    Model directly predicts 100 map elements; fixed by design.
  • decoder stage ratio = 1 non-temporal, 4 temporal
    Selected from Table III as optimal.
  • input image resolution = 704x256
    Input resolution; affects accuracy/speed trade-off.
  • perception range = 60x60m (base), 90x60m (long)
    Range over which map elements are detected.
  • PPDN noise magnitudes
    Rotation, translation, scale, and curvature noise levels are not quantified in the paper; they are free parameters that would need to be tuned.
assumptions (5)
  • domain assumption nuScenes validation mAP is an accepted benchmark for HD map construction.
    The paper evaluates only on nuScenes; results may not transfer to other datasets or sensor setups.
  • domain assumption The reported FPS values on RTX 3090 are measured under comparable conditions across methods.
    FPS comparisons can be affected by framework, batch size, and measurement methodology; the paper states batch size 1 for its own runs but relies on published FPS for baselines.
  • domain assumption SparseDrive provides a valid baseline implementation of sparse perception for maps.
    The paper builds directly on SparseDrive's code and training protocol.
  • domain assumption The auxiliary segmentation head disabled at inference does not affect runtime.
    The paper states it is disabled, so inference FLOPs are unchanged.
  • domain assumption DD3D, nuImages, and ImageNet pretraining are interchangeable initialization strategies whose differences are captured by the reported ablations.
    The paper compares three pretraining choices and selects the best, but does not test the interaction with other components.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SparseMeXT Unlocking the Potential of Sparse Representations for HD Map Construction." pith.science (2026). https://pith.science/paper/4YG2EIHI

@misc{pith2026250508808,
  author       = {Pith},
  title        = {Pith review of: SparseMeXT Unlocking the Potential of Sparse Representations for HD Map Construction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4YG2EIHI}},
  note         = {Machine review of arXiv:2505.08808}
}
read the original abstract

Recent advancements in high-definition \emph{HD} map construction have demonstrated the effectiveness of dense representations, which heavily rely on computationally intensive bird's-eye view \emph{BEV} features. While sparse representations offer a more efficient alternative by avoiding dense BEV processing, existing methods often lag behind due to the lack of tailored designs. These limitations have hindered the competitiveness of sparse representations in online HD map construction. In this work, we systematically revisit and enhance sparse representation techniques, identifying key architectural and algorithmic improvements that bridge the gap with--and ultimately surpass--dense approaches. We introduce a dedicated network architecture optimized for sparse map feature extraction, a sparse-dense segmentation auxiliary task to better leverage geometric and semantic cues, and a denoising module guided by physical priors to refine predictions. Through these enhancements, our method achieves state-of-the-art performance on the nuScenes dataset, significantly advancing HD map construction and centerline detection. Specifically, SparseMeXt-Tiny reaches a mean average precision \emph{mAP} of 55.5% at 32 frames per second \emph{fps}, while SparseMeXt-Base attains 65.2% mAP. Scaling the backbone and decoder further, SparseMeXt-Large achieves an mAP of 68.9% at over 20 fps, establishing a new benchmark for sparse representations in HD map construction. These results underscore the untapped potential of sparse methods, challenging the conventional reliance on dense representations and redefining efficiency-performance trade-offs in the field.

Figures

Figures reproduced from arXiv: 2505.08808 by the authors.

Figure 1
Figure 1. This figure illustrates the trade-off between FPS (frames per second) [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall architecture of SparseMeXt. and SparseAD [32] include detection, tracking and online mapping are unified with general temporal decoders [33], in which multi-sensor features and historical memories are regarded as tokens, and object queries and map queries represent obstacles and road elements in the driving scenario respectively. However, the task of online HD map construc￾tion using sparse architectures rem… view at source ↗
Figure 3
Figure 3. Architecture of the decouple decoder. We independently split [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Architecture of the sparse-dense segmentation head. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Qualitative results of SparseDrive, SparseMeXt, SparseMext LongRange and SparseMeXt Centerline. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. DiffVLA: Vision-Language Guided Diffusion Planning for Autonomous Driving

    cs.AI 2025-05 conditional novelty 4.0 of 10

    DiffVLA integrates VLM guidance, hybrid sparse-dense BEV perception, and a truncated diffusion policy to achieve 45.0 PDMS on the NAVSIM v2 benchmark.

Reference graph

Works this paper leans on

36 extracted references · 12 canonical work pages · cited by 1 Pith paper

  1. [1]

    Loam: Lidar odometry and mapping in real-time

    J. Zhang, S. Singhet al., “Loam: Lidar odometry and mapping in real-time.” inRobotics: Science and systems, vol. 2, no. 9. Berkeley, CA, 2014, pp. 1–9

    2014

  2. [2]

    Lego-loam: Lightweight and ground- optimized lidar odometry and mapping on variable terrain,

    T. Shan and B. Englot, “Lego-loam: Lightweight and ground- optimized lidar odometry and mapping on variable terrain,” in2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018, pp. 4758–4765

    work page 2018

  3. [3]

    Faster- lio: Lightweight tightly coupled lidar-inertial odometry using parallel sparse incremental voxels,

    C. Bai, T. Xiao, Y . Chen, H. Wang, F. Zhang, and X. Gao, “Faster- lio: Lightweight tightly coupled lidar-inertial odometry using parallel sparse incremental voxels,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 4861–4868, 2022

    work page 2022

  4. [4]

    Lego-loam-sc: An improved simultaneous localization and mapping method fusing lego-loam and scan context for underground coalmine,

    G. Xue, J. Wei, R. Li, and J. Cheng, “Lego-loam-sc: An improved simultaneous localization and mapping method fusing lego-loam and scan context for underground coalmine,”Sensors, vol. 22, no. 2, p. 520, 2022

    2022

  5. [5]

    Fast-livo: Fast and tightly-coupled sparse-direct lidar-inertial-visual odometry,

    C. Zheng, Q. Zhu, W. Xu, X. Liu, Q. Guo, and F. Zhang, “Fast-livo: Fast and tightly-coupled sparse-direct lidar-inertial-visual odometry,” in2022 IEEE/RSJ international conference on intelligent robots and systems (IROS). IEEE, 2022, pp. 4003–4009

    work page 2022

  6. [6]

    Persformer: A trans- former architecture for topological machine learning,

    R. Reinauer, M. Caorsi, and N. Berkouk, “Persformer: A trans- former architecture for topological machine learning,”arXiv preprint arXiv:2112.15210, 2021

    arXiv 2021

  7. [7]

    Cross-view transformers for real-time map-view semantic segmentation,

    B. Zhou and P. Kr ¨ahenb¨uhl, “Cross-view transformers for real-time map-view semantic segmentation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 13 760–13 769

    work page 2022

  8. [8]

    Structured bird’s-eye-view traffic scene understanding from onboard images,

    Y . B. Can, A. Liniger, D. P. Paudel, and L. Van Gool, “Structured bird’s-eye-view traffic scene understanding from onboard images,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 15 661–15 670

    2021

Show all 36 references
  1. [9]

    Bevdet: High- performance multi-camera 3d object detection in bird-eye-view,

    J. Huang, G. Huang, Z. Zhu, Y . Ye, and D. Du, “Bevdet: High- performance multi-camera 3d object detection in bird-eye-view,”arXiv preprint arXiv:2112.11790, 2021

    2021 arXiv

  2. [10]

    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,” inEuropean conference on computer vision. Springer, 2022, pp. 1–18

    2022

  3. [11]

    Bevdepth: Acquisition of reliable depth for multi-view 3d object detection,

    Y . Li, Z. Ge, G. Yu, J. Yang, Z. Wang, Y . Shi, J. Sun, and Z. Li, “Bevdepth: Acquisition of reliable depth for multi-view 3d object detection,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 37, no. 2, 2023, pp. 1477–1485

    2023

  4. [12]

    Hdmapnet: An online hd map construction and evaluation framework,

    Q. Li, Y . Wang, Y . Wang, and H. Zhao, “Hdmapnet: An online hd map construction and evaluation framework,” 2022. [Online]. Available: https://arxiv.org/abs/2107.06307

    2022 arXiv

  5. [13]

    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,” 2023. [Online]. Available: https://arxiv.org/abs/2206.08920

    2023 arXiv

  6. [14]

    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,” 2023. [Online]. Available: https://arxiv.org/abs/2208.14437

    2023 arXiv

  7. [15]

    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,” 2024. [Online]. Available: https://arxiv.org/abs/2308.05736

    2024 arXiv

  8. [16]

    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,” 2020. [Online]. Available: https://arxiv.org/abs/2005.12872

    2020 arXiv

  9. [17]

    Sparsedrive: End-to-end autonomous driving via sparse scene representation,

    W. Sun, X. Lin, Y . Shi, C. Zhang, H. Wu, and S. Zheng, “Sparsedrive: End-to-end autonomous driving via sparse scene representation,” 2024. [Online]. Available: https://arxiv.org/abs/2405.19620

    2024 arXiv

  10. [18]

    Persformer: 3d lane detection via perspective transformer and the openlane benchmark,

    L. Chen, C. Sima, Y . Li, Z. Zheng, J. Xu, X. Geng, H. Li, C. He, J. Shi, Y . Qiao, and J. Yan, “Persformer: 3d lane detection via perspective transformer and the openlane benchmark,” 2022. [Online]. Available: https://arxiv.org/abs/2203.11089

    2022 arXiv

  11. [19]

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

    Y . Liu, J. Yan, F. Jia, S. Li, A. Gao, T. Wang, X. Zhang, and J. Sun, “Petrv2: A unified framework for 3d perception from multi-camera images,” 2022. [Online]. Available: https://arxiv.org/abs/2206.01256

    2022 arXiv

  12. [20]

    Dn- detr: Accelerate detr training by introducing query denoising,

    F. Li, H. Zhang, S. Liu, J. Guo, L. M. Ni, and L. Zhang, “Dn- detr: Accelerate detr training by introducing query denoising,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 13 619–13 627

    2022

  13. [21]

    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,” inProceedings of the IEEE/CVF Winter Conference on Appli- cations of Computer Vision, 2024, pp. 7356–7365

    2024

  14. [22]

    Confine keypoint triplets for object detection,

    G. Xu, S. Tang, Z. Yu, and K. Fu, “Confine keypoint triplets for object detection,” in2021 IEEE International Conference on Artificial Intelligence and Industrial Design (AIID). IEEE, 2021, pp. 608–613

    2021

  15. [23]

    Ssd: Single shot multibox detector,

    W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y . Fu, and A. C. Berg, “Ssd: Single shot multibox detector,” inComputer Vision– ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14. Springer, 2016, pp. 21–37

    2016

  16. [24]

    Faster r-cnn: Towards real- time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real- time object detection with region proposal networks,”IEEE transac- tions on pattern analysis and machine intelligence, vol. 39, no. 6, pp. 1137–1149, 2016

    2016

  17. [25]

    Efficientdet: Scalable and efficient object detection,

    M. Tan, R. Pang, and Q. V . Le, “Efficientdet: Scalable and efficient object detection,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 10 781–10 790

    2020

  18. [26]

    Fully convolutional one-stage 3d object detection on lidar range images,

    Z. Tian, X. Chu, X. Wang, X. Wei, and C. Shen, “Fully convolutional one-stage 3d object detection on lidar range images,”Advances in Neural Information Processing Systems, vol. 35, pp. 34 899–34 911, 2022

    2022

  19. [27]

    Semanticformer: Holistic and semantic traffic scene representation for trajectory prediction using knowledge graphs,

    Z. Sun, Z. Wang, L. Halilaj, and J. Luettin, “Semanticformer: Holistic and semantic traffic scene representation for trajectory prediction using knowledge graphs,”IEEE Robotics and Automation Letters, vol. 9, no. 9, pp. 7381–7388, 2024

    2024

  20. [28]

    Detr3d: 3d object detection from multi-view images via 3d-to-2d queries,

    Y . Wang, V . Guizilini, T. Zhang, Y . Wang, H. Zhao, and J. Solomon, “Detr3d: 3d object detection from multi-view images via 3d-to-2d queries,” 2021. [Online]. Available: https://arxiv.org/abs/2110.06922

    2021 arXiv

  21. [29]

    Sparse4d: Multi-view 3d object detection with sparse spatial-temporal fusion,

    X. Lin, T. Lin, Z. Pei, L. Huang, and Z. Su, “Sparse4d: Multi-view 3d object detection with sparse spatial-temporal fusion,”arXiv preprint arXiv:2211.10581, 2022

    2022 arXiv

  22. [30]

    Sparse4d v2: Recurrent temporal fusion with sparse model,

    ——, “Sparse4d v2: Recurrent temporal fusion with sparse model,” arXiv preprint arXiv:2305.14018, 2023

    2023 arXiv

  23. [31]

    Sparse4d v3: Advancing end-to-end 3d detection and tracking,

    X. Lin, Z. Pei, T. Lin, L. Huang, and Z. Su, “Sparse4d v3: Advancing end-to-end 3d detection and tracking,”arXiv preprint arXiv:2311.11722, 2023

    2023 arXiv

  24. [32]

    Sparsead: Sparse query-centric paradigm for efficient end-to-end autonomous driving,

    D. Zhang, G. Wang, R. Zhu, J. Zhao, X. Chen, S. Zhang, J. Gong, Q. Zhou, W. Zhang, N. Wanget al., “Sparsead: Sparse query-centric paradigm for efficient end-to-end autonomous driving,”arXiv preprint arXiv:2404.06892, 2024

    2024 arXiv

  25. [33]

    Exploring object- centric temporal modeling for efficient multi-view 3d object detec- tion,

    S. Wang, Y . Liu, T. Wang, Y . Li, and X. Zhang, “Exploring object- centric temporal modeling for efficient multi-view 3d object detec- tion,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 3621–3631

    2023

  26. [34]

    You should look at all objects,

    Z. Jin, D. Yu, L. Song, Z. Yuan, and L. Yu, “You should look at all objects,” 2022. [Online]. Available: https://arxiv.org/abs/2207.07889

    2022 arXiv

  27. [35]

    You only look one-level feature,

    Q. Chen, Y . Wang, T. Yang, X. Zhang, J. Cheng, and J. Sun, “You only look one-level feature,” 2021. [Online]. Available: https://arxiv.org/abs/2103.09460

    2021 arXiv

  28. [36]

    Mapnext: Revisiting training and scaling practices for online vectorized hd map construction,

    T. Li, “Mapnext: Revisiting training and scaling practices for online vectorized hd map construction,”arXiv preprint arXiv:2401.07323, 2024

    2024 arXiv

Pith tools

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