Pith. sign in

REVIEW 3 major objections 6 minor 2 cited by

MapGS: Generalizable Pretraining and Data Augmentation for Online Mapping via Novel View Synthesis

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Reconstructing one driving dataset with Gaussian splatting and re-rendering it through another vehicle's cameras shrinks the sensor gap that breaks online mapping models; fine-tuning on 25% of target labels then beats full-dataset training.

desk verdict A useful, new data-generation recipe for cross-sensor online mapping, with an honest evaluation and a plausible central claim that would be much stronger with error bars and a clearer SOTA baseline. read the letter →

arxiv 2501.06660 v1 pith:QHGHGRSE submitted 2025-01-11 cs.CV cs.RO

Hengyuan Zhang , David Paz , Yuliang Guo , Xinyu Huang , Henrik I. Christensen , Liu Ren This is my paper · ORCID
classification cs.CVcs.RO
keywords onlinemappingGaussiansplattingnovelviewsynthesissensorconfigurationgeneralizationdatasetaugmentationpretrainingvectorizedHDmapscamera-onlyBEVperception
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 argues that the performance collapse suffered by online mapping models when moved to a vehicle with different cameras is largely a sensor-configuration problem, and that synthetic re-rendering can remove most of it without new labeled data. The authors reconstruct Argoverse 2 driving scenes with Gaussian splatting, then render those scenes through the nuScenes camera rig, producing a synthetic dataset (nuAV2) with images, poses, and map labels all in the target configuration. They report that pretraining on nuAV2 and fine-tuning on only 25% of real nuScenes data outperforms the same model trained from scratch on all of nuScenes (23.0 vs 21.6 mAP), and that joint training raises mAP by 18% over the oracle. Direct cross-sensor generalization improves from 1.2 to 6.3 mAP compared with training on the original Argoverse 2 images. If this holds, existing datasets become reusable assets for new vehicle platforms, reducing the cost of collecting and labeling data per sensor setup.

What carries the argument

The load-bearing mechanism is Street Gaussian scene reconstruction combined with a coordinate-transform chain that re-places the target vehicle inside the reconstructed source scene. StreetGS decomposes each scene into a static background Gaussian field, per-track dynamic vehicle Gaussians that move rigidly along their annotated 3D boxes, and a sky embedding; treating dynamic objects as static along their tracks avoids the view-overfitting seen in alternative dynamic reconstructions such as Periodic Vibration Gaussian. The transform chain maps the source vehicle pose to a target vehicle pose (accounting for different vehicle-frame conventions, such as nuScenes' ground-plane rear axle versus Argoverse 2's raised rear axle), then to each target camera pose, and with target intrinsics renders the images and re-projects the map labels into the target ego-centric frame. The output, nuAV2, is packaged in the nuScenes database format with a virtual LiDAR frame so that it plugs directly into existing training pipelines.

What would settle it

Render a matched experimental pair: take the same physical scenes and compare (a) perceptual or feature distances between nuAV2 renders and real nuScenes images, and (b) the same distances between original Argoverse 2 images and real nuScenes images. The sensor-alignment claim predicts that rendered images are measurably closer to the target configuration than the source images are, and that per-scene fine-tuning gains track this closeness; if renders are not closer, or if the gains do not correlate with render fidelity, then the improvement comes from something other than configuration alignment, such as generic data augmentation or domain regularization.

Watch

Extended reading notes

Core claim

Stated on its own terms, the paper's discovery is that aligning the sensor configuration of training data — not merely adding more data — is what transfers an online mapping model across vehicle platforms. Training a camera-only MapTRv2 on nuAV2, the dataset rendered in the nuScenes configuration, reaches 6.3 mAP on nuScenes evaluation compared with 1.2 mAP for a model trained on the original Argoverse 2 images. Pretraining on nuAV2 improves fine-tuning on every subset size tested (5%, 10%, 25%, 100% of nuScenes), so that a model fine-tuned on 25% of the target data surpasses the oracle trained on 100% of it, and merging nuAV2 with nuScenes yields 25.5 mAP, an 18% relative gain over the oracle. Crucially, pretraining on Argoverse 2 in its native configuration slows convergence and lowers final performance, which isolates the sensor alignment itself as the cause of the gains.

Load-bearing premise

The claim rests on the assumption that images re-rendered from reconstructed Argoverse 2 scenes in nuScenes camera poses are faithful enough to real nuScenes images that training transfers — yet the paper never measures this synthetic-to-real gap, and its own limitation section reports that rendering quality drops sharply when the target viewpoint leaves the recorded trajectory.

Editorial extensions

If this is right

  • A model pretrained on nuAV2 and fine-tuned on 25% of nuScenes reaches 23.0 mAP, surpassing the oracle trained on all of nuScenes (21.6 mAP), while 5% fine-tuning alone reaches 93.1% of oracle performance.
  • Joint training on merged nuAV2 plus nuScenes data lifts mAP to 25.5, an 18% relative improvement over the oracle, so the synthetic data behaves as a genuine augmentation rather than a replacement.
  • Cross-sensor direct generalization improves from 1.2 to 6.3 mAP (a 4.25-times gain) for camera-only models, establishing a camera-only result comparable to the 12.2 mAP of a camera-plus-LiDAR method.
  • Training accelerates: all fine-tuning runs exceed 18 mAP within 2 equivalent epochs, and 5%-data fine-tuning reaches 75% of its final performance by epoch 3.
  • Pretraining in the native (unaligned) Argoverse 2 configuration slows training and reduces final performance, while aligned nuAV2 pretraining speeds it up, showing that alignment, not dataset scale, drives the gain.

Reading between the lines

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

  • The recipe is configuration-agnostic on the rendering side: once a scene is reconstructed, generating data for a third sensor rig (for example a Waymo-style setup) costs only a rendering pass, so the marginal price of a new synthetic dataset collapses to compute and storage rather than data collection.
  • Because the paper does not separate sensor-configuration shift from geographic and domain shift, a cleaner test of the mechanism would render within a single dataset — reconstruct nuScenes and re-render at perturbed camera poses — and check whether similar gains appear; if they do, the effect is calibration-level, not cross-dataset.
  • The reconstruction step still requires LiDAR and 3D track annotations, so the approach currently applies only to datasets that already carry expensive labels; extending it to unlabeled data would need automatic tracking or self-supervised scene decomposition.
  • The 18% joint-training gain and the 2–3 mAP pretraining gains suggest the synthetic images act partly as a regularizer or curriculum; a testable prediction is that higher reconstruction fidelity (measured by render PSNR or feature distance to real target images) should correlate with larger downstream gains.
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

3 major / 6 minor

Summary. The paper proposes MapGS, a data-regeneration paradigm for online HD mapping under cross-sensor configuration. It reconstructs Argoverse 2 scenes with Street Gaussian, renders them in the nuScenes camera configuration (creating a 'nuAV2' dataset), and trains a camera-only MapTRv2 model under direct generalization, pretraining/fine-tuning, and joint-training protocols. The central claims are (i) direct generalization improves from 1.2 to 6.3 mAP over a model trained on original AV2 images, (ii) pretraining on nuAV2 plus fine-tuning on 25% of NUSC yields 23.0 mAP, exceeding the 21.6 mAP Oracle trained on full NUSC, and (iii) joint training gives an 18% relative improvement (25.5 vs. 21.6 mAP). The paper also claims faster convergence and reduced labeling requirements, and it promises public release of code and the nuAV2 dataset.

Significance. If the claims reproduce, the paper makes a useful contribution: a concrete recipe for reusing existing driving data across camera rigs, a large rendered dataset, and a demonstration that novel-view synthesis can serve as data augmentation for vectorized mapping. The paper has strengths: it uses geo-disjoint splits [20], provides qualitative analysis of reconstruction failures and limitations, and is transparent that sensor-configuration and domain gaps are not separated (Sec. 4.6). The computational-cost and storage accounting in the supplement is also valuable. However, the headline quantitative results rest on single training runs without variance estimates, and the sensor-alignment benefit is partly confounded; these issues must be resolved before the superiority claims can be accepted.

major comments (3)
  1. [§4.4–4.5, Tabs. 1–3] The headline numerical claims—25%-fine-tune 23.0 vs. Oracle 21.6 mAP and joint-training 25.5 mAP (an 18% relative improvement)—are reported from a single training run and, for the fine-tuning row, a single random 25% subset of the geo-disjoint split. DETR-style detectors such as MapTRv2 exhibit non-negligible run-to-run variance, and no seeds, error bars, or significance tests are reported anywhere in the paper. The 1.4 mAP margin of Sec. 4.4 is within what run-to-run noise could plausibly produce, and the 18% gain is unreplicated. I request multiple seeds (at least 3) with mean±std, the same seeds across protocols, and a report of the variance of the Oracle baseline. This is load-bearing because the abstract's superiority claims are competition-style quantitative statements.
  2. [§4.1, Tab. 1, §4.6] The direct-generalization comparison (AV2 1.2 mAP vs. nuAV2 6.3 mAP) is confounded: AV2 and NUSC differ not only in camera configuration but also in environment, annotation distributions, and road topology, and the paper itself states in Sec. 4.6 that the evaluation does not separate sensor-configuration generalization from domain generalization. Consequently, the improvement cannot be attributed to sensor alignment alone. To support the central claim, the authors should add a controlled comparison in which the same source scenes are rendered in the source configuration and in the target configuration, or otherwise quantify the domain component (e.g., by training on original AV2 images with aligned labels and identical training budget).
  3. [§3.2 and §4.1] The method's premise is that StreetGS renders NUSC-configuration images faithfully enough for transfer, but no quantitative reconstruction metrics (e.g., PSNR/SSIM/LPIPS) or systematic evaluation of novel-view fidelity are reported. The paper compares PVG and StreetGS only qualitatively (Fig. 3 and Fig. 4), and Sec. 4.6 lists reconstruction quality as a bottleneck. Without measuring the synthetic-to-real gap or its relation to downstream mapping performance, the paper cannot establish that the observed gains come from rendering quality rather than from other properties of the generated data. I request novel-view metrics on held-out AV2 frames and, if possible, an analysis of per-scene render error against downstream mapping mAP.
minor comments (6)
  1. [Throughout] The spacing of 'nuA V2' is inconsistent across the paper (e.g., 'nuA V2' in the abstract vs. 'nuAV2' in Sec. 7.1); please standardize to a single form, such as 'nuAV2'.
  2. [Sec. 4.4] The paper says it randomly samples three subsets of the geo-disjoint NUSC training set, but it does not state the random seed or whether the 5%, 10%, and 25% subsets are nested; please specify this for reproducibility.
  3. [Fig. 3] There is a typo, 'groudtruth', in the figure caption; also, the caption for Fig. 4 is uninformative and should describe what is being shown.
  4. [References] Reference [2] misspells author 'Vora' as 'V ora'; please check all author names for similar typographical errors.
  5. [Sec. 4.5] The claim that 'an epoch in the merged dataset has similar iterations compared to 5 epochs in NUSC' is unclear; please give the exact number of iterations per epoch for NUSC, nuAV2, and the merged dataset.
  6. [Sec. 4.3] The SemVecNet [25] comparison in Tab. 1 uses a different modality (camera+LiDAR) and should be clearly labeled as not like-for-like, ideally moved out of the main comparison table.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the reported gains are empirical comparisons on held-out NUSC validation; the only self-citation (SemVecNet) is a baseline, not a load-bearing premise.

full rationale

This paper's central claims are empirical comparisons on the held-out NUSC validation set: models pretrained on nuAV2 (AV2 scenes rendered into NUSC camera geometry) and fine-tuned on subsets of the geo-disjoint NUSC training split are compared with from-scratch models on the same NUSC training data. The evaluation is not defined in terms of any fitted parameter or training loss; the 25%-subset and 18% improvement numbers are measured rather than derived from the construction of nuAV2. The only in-house citation is SemVecNet [25], used as a baseline and as motivation for the sensor-configuration gap; the paper's own direct-generalization experiment (1.2 mAP from AV2 images vs 6.3 mAP from nuAV2) independently demonstrates the gap, so the self-citation is not load-bearing. The stated limitation that sensor and domain gaps are not separated (Sec. 4.6) and the absence of error bars are statistical/experimental-design concerns, not circularity. No equation reduces to a fitted value, no prediction is defined by its own input, and no uniqueness argument is imported from the authors' prior work.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The central claims rest on the fidelity of the StreetGS reconstructions, the correctness of the sensor calibration transforms, and the transferability of rendered images to real data. No free parameters were fitted; the pipeline uses standard pretraining and fine-tuning hyperparameters reported in the paper.

assumptions (4)
  • domain assumption StreetGS, with LiDAR-initialized static background and track-based dynamic objects, reconstructs AV2 scenes with sufficient fidelity for rendering target-config training images.
    The pipeline depends on reconstruction quality; the paper acknowledges degradation for viewpoints deviating from the trajectory (Sec. 4.1, Fig. 3) and does not provide quantitative reconstruction metrics.
  • domain assumption The geometric transform between AV2 and NUSC vehicle frames and camera extrinsics correctly defines the target sensor configuration.
    Sec. 3.2 defines G_T_VT = G_T_VS * V_S_V_T; any calibration error propagates into misaligned images and labels.
  • domain assumption MapTRv2 without camera embeddings and with the perspective-view segmentation loss removed remains a valid online mapping model for cross-dataset training.
    Supplementary Sec. 7.2 describes these modifications; the paper does not ablate whether the removal affects performance.
  • domain assumption Training on rendered images in the NUSC sensor configuration transfers to real NUSC images, i.e., the synthetic-to-real gap is smaller than the cross-sensor gap.
    This is the central premise of the data-augmentation paradigm; the paper notes in Sec. 4.6 that sensor and domain gaps are not separated.
invented entities (1)
  • Virtual LiDAR frame
    purpose: A synthetic LiDAR TOP frame in nuAV2 used to align map centers between the nuAV2 dataset and NUSC for MapTRv2's map-center reference.
    Described in Supplementary Sec. 7.1; it is a format-level construct with no physical counterpart and is not independently verifiable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MapGS: Generalizable Pretraining and Data Augmentation for Online Mapping via Novel View Synthesis." pith.science (2026). https://pith.science/paper/QHGHGRSE

@misc{pith2026250106660,
  author       = {Pith},
  title        = {Pith review of: MapGS: Generalizable Pretraining and Data Augmentation for Online Mapping via Novel View Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QHGHGRSE}},
  note         = {Machine review of arXiv:2501.06660}
}
read the original abstract

Online mapping reduces the reliance of autonomous vehicles on high-definition (HD) maps, significantly enhancing scalability. However, recent advancements often overlook cross-sensor configuration generalization, leading to performance degradation when models are deployed on vehicles with different camera intrinsics and extrinsics. With the rapid evolution of novel view synthesis methods, we investigate the extent to which these techniques can be leveraged to address the sensor configuration generalization challenge. We propose a novel framework leveraging Gaussian splatting to reconstruct scenes and render camera images in target sensor configurations. The target config sensor data, along with labels mapped to the target config, are used to train online mapping models. Our proposed framework on the nuScenes and Argoverse 2 datasets demonstrates a performance improvement of 18% through effective dataset augmentation, achieves faster convergence and efficient training, and exceeds state-of-the-art performance when using only 25% of the original training data. This enables data reuse and reduces the need for laborious data labeling. Project page at https://henryzhangzhy.github.io/mapgs.

Figures

Figures reproduced from arXiv: 2501.06660 by the authors.

Figure 1
Figure 1. Cross sensor data alignment. Online mapping algorithms struggle when deployed on a vehicle with different sensor configura￾tions and require labeled data with the same sensor configuration. With the source sensor configuration images collected by Argoverse 2 (AV2) [30] data collection vehicles (top row), we propose to leverage Gaussian splatting to render images in the target nuScenes (NUSC) [2] sensor configuration… view at source ↗
Figure 2
Figure 2. MapGS Pipeline. Deploying online mapping models on a different sensor configuration is challenging. MapGS proposes to leverage Street Gaussian (StreetGS) to reconstruct the scene, then render into images in target sensor configuration. We then train a model with this data and labels. Finally, we test the model in target sensor configuration. sky. Ego vehicle masks are added to the ground truth images that capture th… view at source ↗
Figure 3
Figure 3. PVG Distortion. While the reconstruction along the trajectory (top right) has high quality compared to groudtruth (top left), moving the camera away from the trajectory, such as 1 m backwards (bottom left) and 1 m left (bottom right) causes the quality to drop significantly. in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: nuAV2 Examples. nuAV2 renders reconstructed AV2 dataset in the NUSC sensor configuration. In contrast, Street Gaussian (StreetGS) [32] leverages 3D bounding box labels from the dynamic scene to construct a static background and static vehicle Gaussian moving with its t…
Figure 6
Figure 6. Figure 6: 5% NUSC fine-tuning. Models reach over 75% of the final performance with 3 epochs fine-tuning on 5% of the NUSC dataset, showing the effectiveness of pretraining on the nuAV2 dataset. nuAV2 Direct Class Epoch mAP Div. Cross. Bound. Center. mAP 2 6.1 19.5 18.5 31.0 26.9…
Figure 5
Figure 5. Figure 5: Low data fine-tuning. Models pretrained with nuAV2 achieve high performance rapidly, often surpassing their baseline convergence performance, whereas pretraining with AV2 not only slows down training but also reduces overall performance. data is defined as the Oracle. …
Figure 8
Figure 8. Figure 8: MapGS Challenges.. Challenges exist in reconstructing dynamic objects such as pedestrians (top left), and vehicles (top right). Lighting changes and camera auto gain can lead to uneven lighting (bottom left) and floaters (bottom right). configurations. MapGS leverages …
Figure 9
Figure 9. Figure 9: MapTRv2 [18], pretrained on the nuAV2 dataset and fine-tuned with 25% nuScenes (NUSC) [2] dataset, surpasses the Oracle model, trained exclusively on 100% NUSC data. In contrast, pretraining on the Argoverse 2 (AV2) [30] dataset leads to a decline in performance. It is…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. PseudoMapTrainer: Learning Online Mapping without HD Maps

    cs.CV 2025-08 conditional novelty 7.0 of 10

    A method to train online vectorized mapping models from camera images using pseudo-labels built from 2D segmentation and Gaussian splatting, without any ground-truth HD maps.

  2. RGE-GS: Reward-Guided Expansive Driving Scene Reconstruction via Diffusion Priors

    cs.CV 2025-06 conditional novelty 5.0 of 10

    RGE-GS uses a learned pixel-wise reward filter and convergence-aware Gaussian training to improve cross-lane novel view synthesis for driving scenes.

Reference graph

Works this paper leans on

36 extracted references · 31 canonical work pages · cited by 2 Pith papers

  1. [20]

    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 Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22150–22159, 2024. 6, 1

    work page 2024

  2. [1]

    Seitz, and Richard Szeliski

    Sameer Agarwal, Noah Snavely, Ian Simon, Steven M. Seitz, and Richard Szeliski. Building rome in a day. In 2009 IEEE 12th International Conference on Computer Vision , pages 72–79, 2009. 2

    work page 2009

  3. [2]

    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), pages 11618–11628, 2020. 1, 2, 4

    work page 2020

  4. [3]

    Unified domain generalization and adaptation for multi-view 3d object detection, 2024

    Gyusam Chang, Jiwon Lee, Donghyun Kim, Jinkyu Kim, Dongwook Lee, Daehyun Ji, Sujin Jang, and Sangpil Kim. Unified domain generalization and adaptation for multi-view 3d object detection, 2024. 2, 3

    work page 2024

  5. [4]

    Periodic vibration gaussian: Dynamic urban scene reconstruction and real-time rendering

    Yurui Chen, Chun Gu, Junzhe Jiang, Xiatian Zhu, and Li Zhang. Periodic vibration gaussian: Dynamic urban scene reconstruction and real-time rendering. arXiv:2311.18561,

    arXiv

  6. [5]

    Ggs: Generalizable gaussian splatting for lane switching in autonomous driving

    Huasong Han, Kaixuan Zhou, Xiaoxiao Long, Yusen Wang, and Chunxia Xiao. Ggs: Generalizable gaussian splatting for lane switching in autonomous driving. arXiv preprint arXiv:2409.02382, 2024. 3

    arXiv 2024

  7. [6]

    Multiple View Ge- ometry in Computer Vision

    Richard Hartley and Andrew Zisserman. Multiple View Ge- ometry in Computer Vision. Cambridge University Press, 2 edition, 2004. 2

    work page 2004

  8. [7]

    Planning-oriented Autonomous Driving

    Yihan Hu, Jiazhi Yang, Li Chen, Keyu Li, Chonghao Sima, Xizhou Zhu, Siqi Chai, Senyao Du, Tianwei Lin, Wenhai 8 Wang, Lewei Lu, Xiaosong Jia, Qiang Liu, Jifeng Dai, Yu Qiao, and Hongyang Li. Planning-oriented Autonomous Driving. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 17853–17862, 2023. 2

    work page 2023

Show all 36 references
  1. [8]

    Vad: Vectorized scene rep- resentation for efficient autonomous driving

    Bo Jiang, Shaoyu Chen, Qing Xu, Bencheng Liao, Jia- jie Chen, Helong Zhou, Qian Zhang, Wenyu Liu, Chang Huang, and Xinggang Wang. Vad: Vectorized scene rep- resentation for efficient autonomous driving. Proceedings of the IEEE/CVF International Conference on Computer Vision (I...

    2023

  2. [9]

    Polarformer: Multi- camera 3d object detection with polar transformers

    Yanqin Jiang, Li Zhang, Zhenwei Miao, Xiatian Zhu, Jin Gao, Weiming Hu, and Yu-Gang Jiang. Polarformer: Multi- camera 3d object detection with polar transformers. InAAAI,

  3. [10]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk ¨uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. ACM Transactions on Graphics, 42 (4), 2023. 2, 3

    2023

  4. [11]

    Towards viewpoint robustness in bird’s eye view segmentation

    Tzofi Klinghoffer, Jonah Philion, Wenzheng Chen, Or Litany, Zan Gojcic, Jungseock Joo, Ramesh Raskar, Sanja Fidler, and Jose M Alvarez. Towards viewpoint robustness in bird’s eye view segmentation. InInternational Conference on Computer Vision, 2023. 2, 3

    2023

  5. [12]

    Viewformer: Exploring spa- tiotemporal modeling for multi-view 3d occupancy percep- tion via view-guided transformers

    Jinke Li, Xiao He, Chonghua Zhou, Xiaoqiang Cheng, Yang Wen, and Dan Zhang. Viewformer: Exploring spa- tiotemporal modeling for multi-view 3d occupancy percep- tion via view-guided transformers. In Proceedings of the Eu- ropean Conference on Computer Vision (ECCV), pages 90– 1...

    2024

  6. [13]

    Graph-based Topology Reasoning for Driving Scenes

    Tianyu Li, Li Chen, Huijie Wang, Yang Li, Jiazhi Yang, Xiangwei Geng, Shengyin Jiang, Yuting Wang, Hang Xu, Chunjing Xu, Junchi Yan, Ping Luo, and Hongyang Li. Graph-based Topology Reasoning for Driving Scenes. arXiv preprint arXiv:2304.05277, pages 1–12, 2023. 2

    2023 arXiv

  7. [14]

    Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo

    Yinhao Li, Han Bao, Zheng Ge, Jinrong Yang, Jianjian Sun, and Zeming Li. Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo. Pro- ceedings of the AAAI Conference on Artificial Intelligence , 37(2):1486–1494, 2023. 2

    2023

  8. [15]

    Unidrive: Towards universal driving perception across cam- era configurations

    Ye Li, Wenzhao Zheng, Xiaonan Huang, and Kurt Keutzer. Unidrive: Towards universal driving perception across cam- era configurations. arXiv preprint arXiv:2410.13864, 2024. 2, 3

    2024 arXiv

  9. [16]

    BEV- Former: 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. BEV- Former: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers. In Proceedings of the European Conference on Computer Vi- sion (ECCV)...

    2022

  10. [17]

    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 Proceedings of the International Con- ference on Learning Representations (ICLR) , pages 1–18,

  11. [18]

    MapTRv2: An End-to-End Framework for Online Vector- ized 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 Vector- ized HD Map Construction. International Journal of Com- puter Vision, pages 1–17, 2024. 2, 4, 6, 1

    2024

  12. [19]

    Street-gaussians-ns

    LightwheelAI. Street-gaussians-ns. https://github. com / LightwheelAI / street - gaussians - ns ,

  13. [21]

    Ray denoising: Depth-aware hard negative sampling for multi- view 3d object detection

    Feng Liu, Tengteng Huang, Qianjing Zhang, Haotian Yao, Chi Zhang, Fang Wan, Qixiang Ye, and Yanzhao Zhou. Ray denoising: Depth-aware hard negative sampling for multi- view 3d object detection. In Proceedings of the European Conference on Computer Vision (ECCV), 2025. 2

    2025

  14. [22]

    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 Proceedings of the IEEE International Conference on Robotics and Automa...

    2024

  15. [23]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view syn- thesis. In ECCV, 2020. 2

    2020

  16. [24]

    Lift, splat, shoot: Encod- ing images from arbitrary camera rigs by implicitly unpro- jecting to 3d

    Jonah Philion and Sanja Fidler. Lift, splat, shoot: Encod- ing images from arbitrary camera rigs by implicitly unpro- jecting to 3d. In Computer Vision–ECCV 2020: 16th Euro- pean Conference, Glasgow, UK, August 23–28, 2020, Pro- ceedings, Part XIV 16 , pages 194–210. Springer,...

    2020

  17. [25]

    Christensen

    Narayanan Elavathur Ranganatha, Hengyuan Zhang, Shashank Venkatramani, Jing-Yan Liao, and Henrik I. Christensen. Semvecnet: Generalizable vector map gen- eration for arbitrary sensor configurations. In 2024 IEEE Intelligent Vehicles Symposium (IV) , pages 2820–2827,

    2024

  18. [26]

    Scene as occupancy

    Chonghao Sima, Wenwen Tong, Tai Wang, Li Chen, Silei Wu, Hanming Deng, Yi Gu, Lewei Lu, Ping Luo, Dahua Lin, and Hongyang Li. Scene as occupancy. Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023. 2

    2023

  19. [27]

    Scalability in Perception for Autonomous Driving: Waymo Open Dataset

    Pei Sun, Henrik Kretzschmar, Xerxes Dotiwalla, Aur ´elien Chouard, Vijaysai Patnaik, Paul Tsui, James Guo, Yin Zhou, Yuning Chai, Benjamin Caine, Vijay Vasudevan, Wei Han, Jiquan Ngiam, Hang Zhao, Aleksei Timofeev, Scott Et- tinger, Maxim Krivokon, Amy Gao, Aditya Joshi, Yu Zh...

    2020

  20. [28]

    OpenLane-V2: A Topology Reasoning 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, Feng Wen, Hang Xu, Ping Luo, Junchi Yan, 9 Wei Zhang, and Hongyang Li. OpenLane-V2: A Topology Reasoning Benchmark for Unified 3D HD Mapping. In Pro- cee...

    2023

  21. [29]

    To- wards domain generalization for multi-view 3d object detec- tion in bird-eye-view

    Shuo Wang, Xinhai Zhao, Hai-Ming Xu, Zehui Chen, Dameng Yu, Jiahao Chang, Zhen Yang, and Feng Zhao. To- wards domain generalization for multi-view 3d object detec- tion in bird-eye-view. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pa...

    2023

  22. [30]

    Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecasting

    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

  23. [31]

    TopoMLP: A Simple yet Strong Pipeline for Driving Topology Reasoning

    Dongming Wu, Jiahao Chang, Fan Jia, Yingfei Liu, Tiancai Wang, and Jianbing Shen. TopoMLP: A Simple yet Strong Pipeline for Driving Topology Reasoning. In Proceedings of the International Conference on Learning Representations (ICLR), pages 1–12, 2024. 2

    2024

  24. [32]

    Street gaussians: Modeling dynamic urban scenes with gaussian splatting

    Yunzhi Yan, Haotong Lin, Chenxu Zhou, Weijie Wang, Haiyang Sun, Kun Zhan, Xianpeng Lang, Xiaowei Zhou, and Sida Peng. Street gaussians: Modeling dynamic urban scenes with gaussian splatting. In ECCV, 2024. 2, 3, 5

    2024

  25. [33]

    Cross-dataset sensor alignment: Making visual 3d object de- tector generalizable

    Liangtao Zheng, Yicheng Liu, Yue Wang, and Hang Zhao. Cross-dataset sensor alignment: Making visual 3d object de- tector generalizable. In The Conference on Robot Learning,

  26. [34]

    Deformable DETR: Deformable Transform- ers for End-to-End Object Detection

    Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable DETR: Deformable Transform- ers for End-to-End Object Detection. In Proceedings of the International Conference on Learning Representations (ICLR), pages 1–16, 2021. 4, 6 10 MapGS: Generalizable...

    2021

  27. [35]

    Qualitative results Qualitative results for online mapping are presented in Fig. 9. MapTRv2 [18], pretrained on the nuA V2 dataset and fine-tuned with 25% nuScenes (NUSC) [2] dataset, surpasses the Oracle model, trained exclusively on 100% NUSC data. In contrast, pretraining o...

  28. [36]

    nuA V2 in NUSC format To facilitate compatibility with the NUSC development tools, we package the nuA V2 dataset into the NUSC database format

    Additional details of experiments 7.1. nuA V2 in NUSC format To facilitate compatibility with the NUSC development tools, we package the nuA V2 dataset into the NUSC database format. In addition to adopting the NUSC camera setup, we align the vehicle and LiDAR frames in nuA V2...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.