Pith. sign in

REVIEW 4 major objections 5 minor 33 references

M3TR: A Generalist Model for Real-World HD Map Completion

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A single model completes outdated HD maps from any partial prior

desk verdict Useful benchmark and model for HD map completion, but the real-world deployable claim rests on synthetic priors that never contain incorrect map elements. read the letter →

arxiv 2411.10316 v4 pith:EEAFSKV4 submitted 2024-11-15 cs.CV cs.RO

Fabian Immel , Richard Fehler , Frank Bieder , Jan-Hendrik Pauls , Christoph Stiller This is my paper · ORCID
classification cs.CVcs.RO
keywords HDmapcompletionpriorsgeneralistmodelmaskingaugmentationDETRquerydesignmAPCmetricautonomousdriving
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

Self-driving cars rely on high-definition maps that go stale, so the practical question is how to keep them fresh using live sensors while still exploiting the parts that remain valid. M3TR claims that one generalist model, trained by randomly masking semantically coherent classes of ground-truth maps, can handle every kind of map change and perform as well as an ensemble of specialists that each know their change type in advance. On the Argoverse 2 benchmark the generalist matches expert performance, gains up to +4.3 mAPC over the MapEX baseline when a prior is available, and even improves no-prior HD map construction by +1.4 mAP. The paper also contributes a cleaned ground-truth label set, a scenario taxonomy, and a metric, mAPC, that scores only the elements the prior does not supply. If the claim holds, one deployable model can use whatever offline map information exists, without needing to know in advance which elements are outdated.

What carries the argument

The load-bearing mechanism is map masking as an augmentation regime: the whole dataset is used to generate every semantically coherent prior scenario, $M_p = P_p(M_{\mathrm{GT}})$ with $P_p$ masking a full semantic class, so one model sees all change types and gains a four- to five-fold training diversity. Inside the model, prior knowledge enters through a redesigned query encoder: point queries combine a learned prior embedding with prior-derived positional embeddings and reference points, and the one-to-many query set is tiled so every ground-truth repetition also receives map-prior queries (O2MMMP), with prior instances pre-attributed to ground truth before Hungarian matching. The companion metric mAPC evaluates only elements $M_{\mathrm{GT}} \setminus M_p$ absent from the prior, isolating completion performance from trivial prior pass-through.

What would settle it

Run M3TR on real outdated-map data with per-element change labels, such as the Trust but Verify data used only for qualitative motivation, and compare mAPC on genuinely changed elements against the synthetic-scenario mAPC; a drop comparable to the sim-to-real gaps already reported for prior-informed models would falsify the transfer claim.

Watch

Extended reading notes

Core claim

The paper's central claim is that HD map completion should be cast as a generalist problem, not a collection of change-type-specific experts: a single model trained across all prior scenarios (ego lane masked, ego road masked, only road boundaries, only centerlines, no prior) reconstructs the full map as well as or better than specialized models, while using constant memory and no scenario-to-expert oracle. The supporting discovery is that the way prior map elements are fed into the transformer queries matters a lot: replacing zero-padded point embeddings with a learned prior embedding and setting reference points directly from prior geometry (design C), plus tiling prior queries into the one-to-many query set (O2MMMP), accounts for the +4.3 mAPC gain over the MapEX baseline. A third claim is that synthetic masking also acts as data augmentation, improving the no-prior setting even though no architectural change is made for it.

Load-bearing premise

The benchmark and all gains rest on synthetic priors created by masking complete semantic classes from ground-truth maps, and the paper assumes these behave like real outdated HD maps even though no experiment runs on actual map-change data.

Editorial extensions

If this is right

  • A single M3TR generalist can replace an ensemble of prior-scenario experts for deployment, needing about a fifth of the VRAM and no oracle that selects which model to use.
  • Map masking as augmentation transfers to the no-prior task: M3TR improves MapTRv2's no-prior mAP from 46.5 to 47.9 on Argoverse 2 without prior-specific architectural changes.
  • The benchmark's semantically defined scenarios (ego lane, ego road, boundaries, centerlines) and the mAPC metric give future work a standard way to compare completion performance on the elements that actually need online perception.
  • Because M3TR with prior improves class-wise AP for masked elements, downstream planning can trust that elements absent from an outdated map are perceived online rather than silently inherited.

Reading between the lines

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

  • If the augmentation effect is real, the same masking regime could improve other map-related perception tasks such as lane graph construction or change detection without architectural changes; the paper does not test that transfer.
  • The semantic scenario taxonomy suggests a two-stage deployment: a change detector identifies which semantic classes are stale, and the same generalist then completes exactly those classes; the paper explicitly leaves change detection to separate methods.
  • A natural stress test is to combine multiple simultaneously masked classes, since real roads can change in several categories at once; the paper's scenarios mask one class at a time and do not quantify this.
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. Motivated by the fact that offline HD maps become outdated only in parts, the paper proposes M3TR, a MapTRv2-based transformer for HD map completion from a partially valid map prior. The authors contribute (i) a corrected ground-truth label set for Argoverse 2 and nuScenes, (ii) a benchmark with five synthetic prior scenarios obtained by masking entire semantic classes from the ground truth (Eq. 1, Tab. 2), (iii) a prior-aware metric mAPC that scores only elements not contained in the prior, (iv) a point-query and query-set design for injecting prior elements, and (v) a Generalist training regime that mixes all prior scenarios and uses the masks as data augmentation. Experiments on Argoverse 2 and nuScenes report gains over a re-implemented MapEX baseline, with up to +4.3 mAPC for the best expert variant and +1.4 mAP for the generalist without a prior.

Significance. If the central claims hold, the paper would be a useful step toward using offline map priors in online HD map estimation: it provides a clean benchmark, a metric that isolates completion quality, and evidence that a single generalist can replace an ensemble of experts at constant memory cost. The paper has concrete strengths: it provides corrected labels and a geographic split, releases code, compares against an oracle-based expert ensemble, and ablates the query encoding. However, the real-world claim rests on an unvalidated equivalence between synthetic whole-class deletion and actual outdated map priors; the benchmark never exposes the model to incorrect prior geometry. Since the model is trained to pass prior elements through, the reported mAPC gains may not transfer to real map changes. The claims of 'matching expert effectiveness' and 'first real-world deployable model' are therefore stronger than the current evidence.

major comments (4)
  1. [Sec. 3.2, Eq. (1), Supp. Sec. 8] The benchmark derives map priors only by masking elements out of the ground truth, so the model never sees an outdated element that must be corrected. Real outdated maps contain locally invalidated elements that are still present in the prior (e.g., a blocked ego lane or a new bike lane, the motivating examples in Fig. 2), and the model must override them. The statement in Sec. 3.2 that the reconstruction task is indifferent to whether elements are masked synthetically or become masked due to real changes is not supported: synthetic masking makes invalid elements absent, whereas real changes make them present but wrong. This distinction is load-bearing because Supp. Sec. 8 describes a loss that pre-attributes prior queries to ground truth and trains the model to pass prior elements through almost identically, so M3TR has no learned mechanism to update incorrect prior geometry. The proposed mAPC metric (Sec. 3.3) evaluates only elements not in the prior, so the reported +4.3 mAPC cannot detect failure on incorrect prior elements. No experiment uses the real map-change data from Trust but Verify [10], so the abstract's 'first real-world deployable model' claim is not supported by the current evidence.
  2. [Sec. 5.2, Tab. 3, Tab. 6] The claim that the Generalist 'matches the effectiveness of Expert models' is not uniformly supported. On Argoverse 2, the generalist is below the M3TR expert in the MCL scenario (51.2 vs. 55.0 mAPC, Tab. 3) and also slightly below the MapEX baseline there; on nuScenes the generalist is clearly below the M3TR experts for both prior scenarios (MBD 23.8 vs. 26.3, MCL 25.7 vs. 32.5, Tab. 6). The abstract and conclusion state the matching claim without these caveats. The revision should either restrict the claim to the scenarios and datasets where it holds or provide a principled explanation for the nuScenes discrepancy beyond the current dataset-size hypothesis.
  3. [Tab. 4 vs. Tab. 3] The ablation for map masking as augmentation is confounded by data quantity. The 'naive' prior generation in Tab. 4 splits the dataset into disjoint subsets, one per scenario, while the augmentation regime in Tab. 3 trains each scenario on the full dataset. The +0.6 mAPC difference between the two regimes therefore cannot be attributed solely to masking as augmentation; it also reflects an n-fold increase in training data. A matched-data ablation (e.g., subsampling the augmented set so that the naive and augmented regimes see the same number of samples per scenario) is needed to support the specific claim that masking functions as augmentation.
  4. [Sec. 5.2] All quantitative results are single-seed and report only the best checkpoint, with no standard errors or repeated runs. For small differences such as the +1.4 mAP no-prior improvement (Tab. 3) and the 0.2 mAPC difference between point-encoder variants A and B (Tab. 5), this is insufficient to establish that the differences are not due to training noise. Reporting multiple seeds or a clear checkpoint-selection protocol would strengthen the central comparisons.
minor comments (5)
  1. [Sec. 3.3] The definition of mAPC contains a notation error: 'Mp = MGT / Mp' should be the set difference M_GT \ M_p, i.e., the elements missing in the prior.
  2. [Tab. 2 vs. Tab. 3] The scenario names are inconsistent: Tab. 2 uses SEL/SER/SBD/SCL while Tab. 3 uses MEL/MER/MBD/MCL. The notation should be unified.
  3. [Eq. (2)-(3)] The subscript on the prior generator is dropped in Eq. (3): it should be P_p (or P_{p_i}) to match Eq. (2), and the variables S_i and S'_i should be defined before use.
  4. [Sec. 5.2] The training schedule sentence reads '24 / 110 epochs for experts and 54 / 224 epochs for the generalist on Argoverse 2 / nuScenes respectively', which is ambiguous; please clarify which number corresponds to which dataset.
  5. [General] The paper makes strong deployment claims but has no limitations section; a brief discussion of the oracle assumption for invalid-element identification and the use of synthetic priors would be helpful.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: M3TR's reported gains are empirical evaluations on held-out splits with a metric that excludes prior elements; the synthetic-prior limitation is a validity concern, not a circular derivation.

full rationale

The paper's central claims are supported by direct experiments (Tables 3, 5, 6) rather than by definitions that reduce to their own inputs. The completion metric is explicitly designed to exclude prior elements: 'mAPC, which uses not the entire map MGT, but only the map elements Mp = MGT / Mp which are missing in the specific scenario' (Sec. 3.3), so the reported completion gains cannot be inflated by the model passing through prior ground-truth elements. The Generalist-versus-Expert comparison is a held-out evaluation using the external geographic split of [16], and the generalist is evaluated on scenarios it was trained on, which is standard multi-task evaluation rather than a fitted-constant prediction. The training loss does pre-attribute prior instances and trains the model to reproduce prior elements, but this behavior is not presented as a predictive result; the paper explicitly notes that in application one would take prior elements directly from the prior (Supp. Sec. 8 and Table 8 discussion). The main weakness is the synthetic-prior assumption Mp = Pp(MGT) (Eq. 1) and the asserted equivalence between semantic-class masking and real map changes; this is an external-validity limitation, and the paper even acknowledges the oracle assumption: 'The scenarios assume that it is known beforehand which elements are no longer valid, a task for which separate proposed solutions exist' (Sec. 3.2). That is a correctness risk, not a circularity: no equation or fitted parameter makes the claimed +4.3 mAPC equal to an input by construction. The author-group self-citations ([9] for Lanelet2 label generation, [20] for map verification) are methodological and not load-bearing for the central claims. I therefore find no circular step, and assign 1 only to reflect the minor auxiliary self-references.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The method has no new physical entities. Its load-bearing assumptions are the synthetic-prior proxy, the known-invalid oracle, and the sensor sufficiency of the baseline architecture. One architectural hyperparameter (O2M tiling factor) is chosen by hand.

free parameters (1)
  • Number of decoder query repetitions and O2M tiling factor = 70 O2O and 350 O2M queries (factor 5)
    Chosen by hand to match the tiling design; supplement Sec. 9 states total query counts. This architectural choice is part of the proposed query-set design, though its optimality is not derived.
assumptions (3)
  • domain assumption Map priors Mp obtained by masking semantic classes from ground truth MGT (Eq. 1) are representative of real outdated HD maps.
    The whole benchmark and evaluation rely on this; real map-change data is used only as motivation, not for evaluation.
  • domain assumption The set of invalid prior elements is known beforehand for each scenario.
    Sec. 3.2 explicitly assumes this and delegates change detection to separate methods; deployment would need this oracle in practice.
  • domain assumption A ResNet50 image backbone plus BEV transformation provides enough information to reconstruct masked map elements.
    This is the standard MapTRv2 assumption inherited without revalidation on real change scenarios.

how reviews work

0 comments
Cite this review

Pith. "Pith review of M3TR: A Generalist Model for Real-World HD Map Completion." pith.science (2026). https://pith.science/paper/EEAFSKV4

@misc{pith2026241110316,
  author       = {Pith},
  title        = {Pith review of: M3TR: A Generalist Model for Real-World HD Map Completion},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EEAFSKV4}},
  note         = {Machine review of arXiv:2411.10316}
}
read the original abstract

Autonomous vehicles rely on HD maps for their operation, but offline HD maps eventually become outdated. For this reason, online HD map construction methods use live sensor data to infer map information instead. Research on real map changes shows that oftentimes entire parts of an HD map remain unchanged and can be used as a prior. We therefore introduce M3TR (Multi-Masking Map Transformer), a generalist approach for HD map completion both with and without offline HD map priors. As a necessary foundation, we address shortcomings in ground truth labels for Argoverse 2 and nuScenes and propose the first comprehensive benchmark for HD map completion. Unlike existing models that specialize in a single kind of map change, which is unrealistic for deployment, our Generalist model handles all kinds of changes, matching the effectiveness of Expert models. With our map masking as augmentation regime, we can even achieve a +1.4 mAP improvement without a prior. Finally, by fully utilizing prior HD map elements and optimizing query designs, M3TR outperforms existing methods by +4.3 mAP while being the first real-world deployable model for offline HD map priors. Code is available at https://github.com/immel-f/m3tr

Figures

Figures reproduced from arXiv: 2411.10316 by the authors.

Figure 1
Figure 1. Overview of the model architecture of M3TR and the investigated point query encoder designs. For our evaluated task of HD map [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Visualization of map changes from [10], with the out￾dated map reprojected into the camera image. Real map changes can easily be translated into the proposed map prior scenarios. become blocked, resulting in invalidated elements akin to SEL. In Fig. 2b, a bike lane is added that causes the ego road to become invalid, similar to SER. The scenarios assume that it is known beforehand which elements are no longer valid,… view at source ↗
Figure 3
Figure 3. Visualization of different detection query set designs [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (9 more)
Figure 5
Figure 5. Figure 5: Visualization of previous expert models vs. the [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Example of the M3TR generalist model on the same sample from Argoverse 2 with different map priors. The more information [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Comparisons of commonly used labels versus our pro [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Visualization of the pre-attribution matching strategy [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Commonly used ground truth and our new ground truth on Argoverse 2 reprojected into the associated camera images. Many [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: Qualitative examples comparing the M3TR Generalist without any prior with MapTRv2 [14] on Argoverse 2. M3TR shows improved performance even without any prior due to our proposed masking as augmentation [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Qualitative examples comparing the M3TR Generalist with the respective MapEX [23] Experts on Argoverse 2. 5 [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]
Figure 12
Figure 12. Figure 12: More qualitative examples comparing the M3TR [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 13
Figure 13. Figure 13: Qualitative examples comparing the M3TR Experts with the respective MapEX [23] Experts on nuScenes. 7 [PITH_FULL_IMAGE:figures/full_fig_p017_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 22 canonical work pages

  1. [10]

    Trust, but verify: Cross- modality fusion for hd map change detection

    John Lambert and James Hays. Trust, but verify: Cross- modality fusion for hd map change detection. In Proceed- ings of the Neural Information Processing Systems Track on Datasets and Benchmarks. Curran, 2021. 1, 3, 4

    work page 2021

  2. [1]

    ASAM OpenDRIVE 1.8.0 Specification , 2023

    ASAM e.V . ASAM OpenDRIVE 1.8.0 Specification , 2023. Published November 22, 2023. 1

    work page 2023

  3. [2]

    Bateman, Ning Xu, H

    Samuel M. Bateman, Ning Xu, H. Charles Zhao, Yael Ben Shalom, Vince Gong, Greg Long, and Will Maddern. Exploring real world map change generalization of prior- informed hd map prediction models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 4568–4578, 2024. 3, 4

    work page 2024

  4. [3]

    Lang, Sourabh V ora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Gi- ancarlo Baldan, and Oscar Beijbom

    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 (CVPR), 2020. 2

    2020

  5. [4]

    End-to- end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to- end object detection with transformers. In European confer- ence on computer vision, pages 213–229. Springer, 2020. 2

    2020

  6. [5]

    Maptracker: Tracking with strided memory fusion for consistent vector hd mapping

    Jiacheng Chen, Yuefan Wu, Jiaqi Tan, Hang Ma, and Yasu- taka Furukawa. Maptracker: Tracking with strided memory fusion for consistent vector hd mapping. In Computer Vision – ECCV 2024, pages 90–107, Cham, 2025. Springer Nature Switzerland. 1, 2, 3, 4

    work page 2024

  7. [6]

    Efficient and robust 2d-to-bev representation learning via geometry-guided ker- nel transformer

    Shaoyu Chen, Tianheng Cheng, Xinggang Wang, Wenming Meng, Qian Zhang, and Wenyu Liu. Efficient and robust 2d-to-bev representation learning via geometry-guided ker- nel transformer. arXiv preprint arXiv:2206.04584, 2022. 2

    arXiv 2022

  8. [7]

    Mask2Map: Vectorized HD Map Construction Using Bird's Eye View Segmentation Masks

    Sehwan Choi, Jungho Kim, Hongjae Shin, and Jun Won Choi. Mask2map: Vectorized hd map construction us- ing bird’s eye view segmentation masks. arXiv preprint arXiv:2407.13517, 2024. 2

    work page Pith review arXiv 2024

Show all 33 references
  1. [8]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016. 6

    2016

  2. [9]

    Generation of training data from hd maps in the lanelet2 framework

    Fabian Immel, Richard Fehler, Frank Bieder, and Christoph Stiller. Generation of training data from hd maps in the lanelet2 framework. arXiv preprint arXiv:2407.17409, 2024. 1

    2024 arXiv

  3. [11]

    Hdmapnet: An online hd map construction and evaluation framework

    Qi Li, Yue Wang, Yilun Wang, and Hang Zhao. Hdmapnet: An online hd map construction and evaluation framework. In 2022 International Conference on Robotics and Automation (ICRA), pages 4628–4634, 2022. 1

    2022

  4. [12]

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

    Zhiqi Li, Wenhai Wang, Hongyang Li, Enze Xie, Chong- hao Sima, Tong Lu, Yu Qiao, and Jifeng Dai. Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers. In European con- ference on computer vision, pages 1–18. Springer, 2022. 2

    2022

  5. [13]

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

    Bencheng Liao, Shaoyu Chen, Xinggang Wang, Tianheng Cheng, Qian Zhang, Wenyu Liu, and Chang Huang. Maptr: Structured modeling and learning for online vectorized hd map construction. In The Eleventh International Conference on Learning Representations, 2022. 1, 2, 6, 4

    2022

  6. [14]

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

    Bencheng Liao, Shaoyu Chen, Yunchi Zhang, Bo Jiang, Qian Zhang, Wenyu Liu, Chang Huang, and Xinggang Wang. Maptrv2: An end-to-end framework for online vectorized hd map construction. International Journal of Computer Vision,

  7. [15]

    Lane graph as path: Continuity-preserving path-wise modeling for online lane graph construction

    Bencheng Liao, Shaoyu Chen, Bo Jiang, Tianheng Cheng, Qian Zhang, Wenyu Liu, Chang Huang, and Xinggang Wang. Lane graph as path: Continuity-preserving path-wise modeling for online lane graph construction. In Computer Vision – ECCV 2024, pages 334–351, Cham, 2025. Springer Nat...

    2024

  8. [16]

    Localization is all you evaluate: Data leakage in online mapping datasets and how to fix it

    Adam Lilja, Junsheng Fu, Erik Stenborg, and Lars Ham- marstrand. Localization is all you evaluate: Data leakage in online mapping datasets and how to fix it. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pat- tern Recognition (CVPR), pages 22150–22159, 2024....

    2024

  9. [17]

    VectorMapNet: End-to-end vectorized HD map learning

    Yicheng Liu, Tianyuan Yuan, Yue Wang, Yilun Wang, and Hang Zhao. VectorMapNet: End-to-end vectorized HD map learning. In Proceedings of the 40th International Con- ference on Machine Learning, pages 22352–22369. PMLR,

  10. [18]

    Augmenting lane perception and topology understanding with standard definition navigation maps

    Katie Z Luo, Xinshuo Weng, Yan Wang, Shuang Wu, Jie Li, Kilian Q Weinberger, Yue Wang, and Marco Pavone. Augmenting lane perception and topology understanding with standard definition navigation maps. arXiv preprint arXiv:2311.04079, 2023. 4

    2023 arXiv

  11. [19]

    Aug- menting lane perception and topology understanding with standard definition navigation maps

    Katie Z Luo, Xinshuo Weng, Yan Wang, Shuang Wu, Jie Li, Kilian Q Weinberger, Yue Wang, and Marco Pavone. Aug- menting lane perception and topology understanding with standard definition navigation maps. In 2024 IEEE Inter- national Conference on Robotics and Automation (ICRA) ...

    2024

  12. [20]

    Hd map verification without ac- curate localization prior using spatio-semantic 1d signals

    Jan-Hendrik Pauls, Tobias Strauss, Carsten Hasberg, Martin Lauer, and Christoph Stiller. Hd map verification without ac- curate localization prior using spatio-semantic 1d signals. In 2020 IEEE Intelligent Vehicles Symposium (IV), pages 680– 686, 2020. 4

    2020

  13. [21]

    Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unproject- 9 ing to 3d

    Jonah Philion and Sanja Fidler. Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unproject- 9 ing to 3d. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceed- ings, Part XIV 16, pages 194–210. Springer, 2020. 2

    2020

  14. [22]

    Lanelet2: A high-definition map framework for the future of automated driving

    Fabian Poggenhans, Jan-Hendrik Pauls, Johannes Janoso- vits, Stefan Orf, Maximilian Naumann, Florian Kuhnt, and Matthias Mayr. Lanelet2: A high-definition map framework for the future of automated driving. In Proc. IEEE Intell. Trans. Syst. Conf., Hawaii, USA, 2018. 1

    2018

  15. [23]

    Mind the map! accounting for existing map information when estimating online hdmaps from sensor data

    R ´emy Sun, Li Yang, Diane Lingrand, and Fr´ed´eric Precioso. Mind the map! accounting for existing map information when estimating online hdmaps from sensor data. arXiv preprint arXiv:2311.10517, 2024. 1, 2, 3, 4, 5, 6, 7, 8

    2024 arXiv

  16. [24]

    Openlane-v2: A topology reason- ing benchmark for unified 3d hd mapping

    Huijie Wang, Tianyu Li, Yang Li, Li Chen, Chonghao Sima, Zhenbo Liu, Bangjun Wang, Peijin Jia, Yuting Wang, Shengyin Jiang, et al. Openlane-v2: A topology reason- ing benchmark for unified 3d hd mapping. In Thirty- seventh Conference on Neural Information Processing Sys- tems ...

    2023

  17. [25]

    Exelmap: Explainable element-based hd-map change detection and update

    Lena Wild, Ludvig Ericson, Rafael Valencia, and Patric Jens- felt. Exelmap: Explainable element-based hd-map change detection and update. In Proceedings of the ECCV 2024 2nd Workshop on Vision-Centric Autonomous Driving (VCAD) ,

    2024

  18. [26]

    Argoverse 2: Next generation datasets for self-driving perception and fore- casting

    Benjamin Wilson, William Qi, Tanmay Agarwal, John Lam- bert, Jagjeet Singh, Siddhesh Khandelwal, Bowen Pan, Rat- nesh Kumar, Andrew Hartnett, Jhony Kaesemodel Pontes, Deva Ramanan, Peter Carr, and James Hays. Argoverse 2: Next generation datasets for self-driving perception an...

    2021

  19. [27]

    Streammapnet: Streaming mapping network for vectorized online hd map construction

    Tianyuan Yuan, Yicheng Liu, Yue Wang, Yilun Wang, and Hang Zhao. Streammapnet: Streaming mapping network for vectorized online hd map construction. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 7356–7365, 2024. 2, 3, 6, 1, 4

    2024

  20. [28]

    Driving with prior maps: Unified vector prior encoding for autonomous vehicle mapping

    Shuang Zeng, Xinyuan Chang, Xinran Liu, Zheng Pan, and Xing Wei. Driving with prior maps: Unified vector prior encoding for autonomous vehicle mapping. arXiv preprint arXiv:2409.05352, 2024. 3, 4, 1

    2024

  21. [29]

    Himap: Hybrid represen- tation learning for end-to-end vectorized hd map construc- tion

    Yi Zhou, Hui Zhang, Jiaqian Yu, Yifan Yang, Sangil Jung, Seung-In Park, and ByungIn Yoo. Himap: Hybrid represen- tation learning for end-to-end vectorized hd map construc- tion. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), page...

    2024

  22. [31]

    Detailed Description of Changes and Motivation Compared with semantic HD planning maps used in map- based autonomous driving stacks using e.g

    Remaining Details of Novel Ground Truth This section presents more details of our novel ground truth, namely a more detailed description of the changes and their motivation, qualitative examples of commonly used and new labels and a comparison of label features on the nuScenes...

  23. [32]

    4.1, we follow MapEX [23] for the training loss, including the pre-attribution of map prior in- stances during assignment

    Details of Loss Pre-Attribution As mentioned in Sec. 4.1, we follow MapEX [23] for the training loss, including the pre-attribution of map prior in- stances during assignment. A visualization of the pre- attribution strategy can be seen in Fig. 8. All ground truth instances gi...

  24. [33]

    Additional Evaluation Results and Details This section presents additional evaluation results on Argo- verse 2 and qualitative examples to complement Sec. 5. Remaining Implementation Details As shown in Fig. 3, the total number of O2M queries is cho- sen as a multiple of the g...

  25. [2024]

    1, 2, 3, 4, 5, 6, 7, 8

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.