Pith. sign in

REVIEW 4 major objections 4 minor 2 cited by

LDMapNet-U: An End-to-End System for City-Scale Lane-Level Map Updating

T0 review · 4 major / 4 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 Lane-level map updating—normally a multi-stage, labor-heavy pipeline—can be compressed into a single end-to-end model that takes recent bird's-eye-view images and the existing map, and returns both the updated vectorized map and per-lane…

desk verdict A credible industrial system for lane-level map updating with a real evaluation gap: the headline gains over baselines are not controlled for the historical-map input. read the letter →

arxiv 2501.02763 v2 pith:4OW5RNKI submitted 2025-01-06 cs.CV

Deguo Xia , Weiming Zhang , Xiyan Liu , Wei Zhang , Chenting Gong , Xiao Tan , Jizhou Huang , Mengmeng Yang
show 1 more author
Diange Yang
This is my paper · ORCID
classification cs.CV
keywords Lane-levelmapupdatingEnd-to-endPriorChangedetectionVectorizedBEVimageAutonomousdriving
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

This paper claims that lane-level map updating—typically a labor-intensive sequence of map construction, human change detection, and manual redrawing—can be recast as a single end-to-end generation task. The proposed system, LDMapNet-U, takes a recent bird's-eye-view (BEV) image and the existing vectorized map as input, and directly outputs the updated vectorized map together with per-instance change labels (no change, style change, addition, deletion). The authors argue that this removes the error accumulation and manual post-processing of multi-stage baselines, and they report strong recall and precision gains over prior methods on a twelve-city real-world dataset. The strongest evidence is operational: the system has been running in production at Baidu Maps since April 2024, updating lane-level maps for more than 360 cities and shortening the update cycle from quarterly to weekly.

What carries the argument

Two new modules carry the argument. The Prior-Map Encoding (PME) module converts each historical lane instance into positional and semantic embeddings, refines them with self-attention layers, and fuses the resulting map embeddings into the image feature map via cross-attention, giving the model the historical geometry and style as prior context. The Instance Change Prediction (ICP) module builds an affinity matrix between predicted instance features and historical instance features, applies Hungarian matching to produce a one-to-one association matrix, and then derives the four change categories (no change, style change, addition, deletion) directly from matched and unmatched instances and style comparisons.

What would settle it

Independently re-annotate a random sample of about 500 scenes from the LD-U test set with different human experts and compute agreement with the paper's ground-truth change labels; if inter-annotator agreement on change type falls below the reported change recall and precision, the metrics overstate what the model can be trusted to do in the field.

Watch

Extended reading notes

Core claim

The central claim is that historical map data, when encoded as a learnable prior, makes both map generation and change detection more accurate than using the BEV image alone. LDMapNet-U fuses the historical map into the BEV feature space via a Prior-Map Encoding (PME) module, and predicts change labels by learning an association matrix between predicted instances and historical instances through an Instance Change Prediction (ICP) module. With these modules, the model simultaneously predicts vectorized elements and their change types in one forward pass, reaching an average recall of 92.62 percent on the larger training set and change recall around 91.5 percent, numbers the paper reports as clear improvements over the three-stage baselines.

Load-bearing premise

The evaluation assumes that the historical map and the BEV image are accurately aligned and that the labeled 'updated map' ground truth is correct, because the training and test labels come from automated instance matching later refined by human experts; if the historical map contains errors, the model learns to reproduce them.

Editorial extensions

If this is right

  • If the deployed claims hold, lane-level maps for an entire country can be kept weekly-fresh without per-edit manual annotation.
  • Change labels are produced end-to-end from the same features that generate the map, so the updated geometry and the change report are mutually consistent.
  • Ablation results show that fusing historical map embeddings at the BEV feature level rather than at the decoder queries yields a 2.32 percent gain in map recall and a 2.9 percent gain in change recall, indicating where prior information matters most.
  • On a larger 1.5 million-sample training set, the model scales to 92.62 percent average recall, suggesting the approach continues to benefit from more data.

Reading between the lines

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

  • A likely implicit prerequisite is that the historical map itself is mostly correct; the method may faithfully reproduce any systematic errors already present in the prior map, and the reported gains over no-prior baselines partly reflect the prior's share of correct geometry.
  • The 95 percent cost reduction and weekly-cycle claims are operationally measured at Baidu, but the paper does not publish a standardized external benchmark for map updating, so cold comparison with other published methods on the same data is limited.
  • The dependency on decimeter-level localization for the crowdsourced BEV collection (stated in Section 5) suggests the approach may not transfer to low-cost fleets without a localization or co-registration step.
  • A testable extension would apply LDMapNet-U with a deliberately corrupted or outdated historical map (for example, a one-year-old map) and measure how much of the performance gain persists, which would separate the prior's contribution from image-based perception.
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

4 major / 4 minor

Summary. LDMapNet-U proposes an end-to-end system for city-scale lane-level map updating. Given a BEV image and the corresponding historical vectorized map, the model simultaneously outputs an updated vectorized map and per-instance change labels (no change, style change, instance addition, instance deletion). The architecture introduces a Prior-Map Encoding (PME) module that encodes historical map elements and fuses them with BEV features, and an Instance Change Prediction (ICP) module that builds an association matrix between predicted and historical instances, followed by a Hungarian matching step and a rule-based derivation of change categories. The model is trained with a multi-task loss combining map-element prediction and association classification, and is evaluated on a large real-world dataset (LD-U and LD-U-L) against MapTR, GeMap, and DuMapNet. The paper claims production deployment at Baidu Maps since April 2024, supporting weekly updates for over 360 cities.

Significance. If the results are taken at face value, the paper describes a practically impactful industrial system: an end-to-end updater that shortens map refresh cycles and is deployed at national scale. The paper's strengths are the large real-world evaluation dataset, the explicit treatment of lane-level change categories, the ablation studies for fusion and association modules, and the credible production-deployment narrative. The architectural idea of conditioning vectorized map generation on a historical map prior and jointly predicting changes is sensible and likely to be of interest to both the academic and industrial communities. However, the quantitative comparison to prior art is confounded by the fact that the baselines do not receive the historical map input; most of the reported gain is attributable to the privileged prior rather than to the end-to-end paradigm itself. The change-detection comparison is also weakened by an undescribed post-processing procedure for the baselines. These issues are load-bearing for the paper's central claim of 'superiority,' so the evaluation needs to be substantially reworked.

major comments (4)
  1. [§3.2, Table 2] The comparison to MapTR, GeMap, and DuMapNet is not input-equivalent: LDMapNet-U additionally receives the historical vectorized map V_Io, while the baselines receive only the BEV image. On mostly unchanged roads, a model that simply copies the historical map would achieve high recall without performing any update, so the reported gains (+7.69 to +15.76 percentage points) do not isolate the proposed end-to-end paradigm from the informational advantage of the historical-map prior. Please add a 'copy historical map' baseline and/or adapt the baselines to also consume the historical map (e.g., via the same PME module) for a controlled comparison.
  2. [§3.2, Table 4] The 'w/o Fusion' ablation is not a no-prior control: as shown in Figure 3a, historical map embeddings are still fed into the Map Association module, and the ICP module still uses them to build the association matrix. Its mR@P=80% of 75.37 is close to DuMapNet's 75.29, suggesting that most of the improvement over baselines comes from the historical-map prior rather than from the end-to-end association learning. A control that removes the historical map entirely (or a baseline that uses it in a simpler way) is needed to support the claim that the proposed architecture is the source of the gains.
  3. [§2.5, §3.2, Table 3] The change-detection comparison is difficult to interpret because the baselines' change labels are produced by an undescribed 'existing industrial-grade post-processing logic' whose thresholds are never given; the +9.18 to +13.84 percentage-point recall gains could reflect a deliberately conservative post-processor. In addition, the change categories of LDMapNet-U are not directly learned: they are derived by a deterministic rule from the association matrix and a style-consistency check (Section 2.5), and the 'filtering' step introduces a confidence threshold. Please specify the post-processing thresholds for all methods and either predict change types directly or describe the rule-based derivation as part of the inference procedure.
  4. [§2.6, Eq. (4)] The training loss L_c supervises only the association matrix A_ij, not the change-type labels (style change, addition, deletion, no change). The change types are inferred at inference time by combining the learned associations with the separately predicted styles via the rules in Section 2.5. This is a legitimate design, but the paper should state explicitly that the change-type predictor is a fixed rule rather than a learned classifier, and the term 'end-to-end change prediction' should be qualified accordingly.
minor comments (4)
  1. [§3.1, Table 1] The validation row lists 15,884 images but no mileage value, while the 'All' row reports 9,890 km; please clarify whether validation kilometers are included and fix the column alignment.
  2. [§2.3, §3.1, §3.3] There are several typos: 'structurre' in Section 2.3, 'piror' in Section 3.1, 'Futher' in Section 3.3, and 'includs' in Section 2.3. Please proofread the text.
  3. [Figure 2c] The dimensions of matrices A and M are written as R^{!×#}, which appears to be a placeholder; please replace with the actual dimensions, R^{N×M} for both.
  4. [§4.2] The related-work section on map change detection is brief and omits several recent learning-based HD map change detection methods; consider expanding it to better position the contribution.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the historical map is an input feature, change labels are externally supervised, and self-citations to DuMapNet are not load-bearing.

full rationale

The paper's derivation chain is self-contained rather than circular. The change-detection output is defined in Section 2.5 as a deterministic function of the learned association matrix M and style comparison: 'No change: M_ij = 1 and the styles ... are consistent; Style change: M_ij = 1 and the styles ... differ; Instance addition: M_i = 0 ...; Instance deletion: M_j = 0.' The ground-truth change labels are produced independently in Section 2.1 by an 'automated stage employed instance matching and comparison' followed by 'human expert refinement'; the model is trained against these external labels via the association classification loss in Eq. (4). No fitted parameter is renamed as a prediction, and no output is defined in terms of itself. The historical map V_Io is an input feature, not a learned parameter; the large gains over image-only baselines in Table 2 are explicitly attributed by the authors to the prior-map input ('This is reasonable since LDMapNet-U sufficiently learns from the input historical vectorized map data'), which is a controlled-input/fairness concern rather than circularity. The self-citations to DuMapNet [32] are used for loss design, BEV resolution conventions, and as a baseline; DuMapNet is an independently published system, and these citations are not invoked as a uniqueness theorem or as the sole justification for the core claim. The baseline comparison for change detection uses an undescribed industrial post-processing logic, which limits reproducibility, but that is an evaluation-transparency issue, not a circular reduction. Overall, no load-bearing step reduces to its own inputs by construction.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The paper relies on standard deep learning tools and on domain-specific assumptions about data quality, alignment, and representation. The only hand-set parameters are architectural and loss hyperparameters; there are no newly invented physical entities.

free parameters (6)
  • loss weight alpha = 1
    Equation (2); chosen by hand, balances map element loss and change loss.
  • loss weight beta = 1
    Equation (2); chosen by hand.
  • number of instance queries = 50
    Section 3.1 Implementation Details; set by the authors.
  • number of point queries per element = 50
    Section 2.3 and 3.1; fixed point set size for each map element.
  • number of decoder layers = 6
    Section 3.1.
  • learning rate = 6e-4
    Section 3.1 with cosine decay.
assumptions (4)
  • domain assumption Bird's-eye-view images accurately reflect current road geometry and are aligned with historical map coordinates.
    Section 2.1 builds sample pairs from BEV images and historical maps; misalignment would corrupt both training and evaluation.
  • domain assumption Ground-truth updated maps and change labels produced by automated matching plus human refinement are correct.
    Section 2.1; the model trains and is evaluated against these labels, so labeling errors directly bias reported metrics.
  • domain assumption High-precision crowdsourced data with decimeter-level or better localization is available.
    Section 5 states the method relies on this; lower-precision data is left as an open problem.
  • ad hoc to paper Every lane element can be represented as a fixed-size set of points.
    Section 2.3 fixes Np=50 points per element; long or complex geometries may be truncated, a modeling choice not independently justified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of LDMapNet-U: An End-to-End System for City-Scale Lane-Level Map Updating." pith.science (2026). https://pith.science/paper/4OW5RNKI

@misc{pith2026250102763,
  author       = {Pith},
  title        = {Pith review of: LDMapNet-U: An End-to-End System for City-Scale Lane-Level Map Updating},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4OW5RNKI}},
  note         = {Machine review of arXiv:2501.02763}
}
read the original abstract

An up-to-date city-scale lane-level map is an indispensable infrastructure and a key enabling technology for ensuring the safety and user experience of autonomous driving systems. In industrial scenarios, reliance on manual annotation for map updates creates a critical bottleneck. Lane-level updates require precise change information and must ensure consistency with adjacent data while adhering to strict standards. Traditional methods utilize a three-stage approach-construction, change detection, and updating-which often necessitates manual verification due to accuracy limitations. This results in labor-intensive processes and hampers timely updates. To address these challenges, we propose LDMapNet-U, which implements a new end-to-end paradigm for city-scale lane-level map updating. By reconceptualizing the update task as an end-to-end map generation process grounded in historical map data, we introduce a paradigm shift in map updating that simultaneously generates vectorized maps and change information. To achieve this, a Prior-Map Encoding (PME) module is introduced to effectively encode historical maps, serving as a critical reference for detecting changes. Additionally, we incorporate a novel Instance Change Prediction (ICP) module that learns to predict associations with historical maps. Consequently, LDMapNet-U simultaneously achieves vectorized map element generation and change detection. To demonstrate the superiority and effectiveness of LDMapNet-U, extensive experiments are conducted using large-scale real-world datasets. In addition, LDMapNet-U has been successfully deployed in production at Baidu Maps since April 2024, supporting map updating for over 360 cities and significantly shortening the update cycle from quarterly to weekly. The updated maps serve hundreds of millions of users and are integrated into the autonomous driving systems of several leading vehicle companies.

Figures

Figures reproduced from arXiv: 2501.02763 by the authors.

Figure 1
Figure 1. LDMapNet-U presents an end-to-end automated [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (a) Overall architecture of LDMapNet-U. (b) Our proposed Prior-Map Encoding (PME) module. (c) Our proposed [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Different fusion methods for historical map embeddings. (a) w/o Fusion indicates that the historical map embeddings [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative comparisons of our model with several state-of-the-art models. Instances addition, instance deletion, style [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

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. MagicRoad: Semantic-Aware 3D Road Surface Reconstruction via Obstacle Inpainting

    cs.CV 2025-07 reject novelty 5.0 of 10

    MagicRoad combines video inpainting, semantic color harmonization, and 2D Gaussian surfels to reconstruct clean bird's eye view road surfaces from driving video.

Reference graph

Works this paper leans on

39 extracted references · 29 canonical work pages · cited by 2 Pith papers

  1. [1]

    Shuhui Bu, Qing Li, Pengcheng Han, Pengyu Leng, and Ke Li. 2020. Mask-CDNet: A mask based pixel change detection network. Neurocomputing 378 (2020), 166–178

    work page 2020

  2. [2]

    Tom Bu, Christoph Mertz, and John Dolan. 2023. Toward map updates with crosswalk change detection using a monocular bus camera. In2023 IEEE Intelligent Vehicles Symposium (IV). IEEE, 1–8

    work page 2023

  3. [3]

    Pengxin Chen, Xiaoqi Jiang, Yingjun Zhang, Jiahao Tan, and Rong Jiang. 2024. MapCVV: On-cloud Map Construction Using Crowdsourcing Visual Vectorized Elements towards Autonomous Driving. IEEE Robotics and Automation Letters (2024)

    work page 2024

  4. [4]

    Wenjie Ding, Limeng Qiao, Xi Qiu, and Chi Zhang. 2023. Pivotnet: Vectorized pivot learning for end-to-end hd map construction. InProceedings of the IEEE/CVF International Conference on Computer Vision . 3672–3682

    work page 2023

  5. [5]

    Xiaomin Fang, Jizhou Huang, Fan Wang, Lihang Liu, Yibo Sun, and Haifeng Wang. 2021. SSML: Self-Supervised Meta-Learner for En Route Travel Time Estimation at Baidu Maps. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining . 2840–2848

    work page 2021

  6. [6]

    Xiaomin Fang, Jizhou Huang, Fan Wang, Lingke Zeng, Haijin Liang, and Haifeng Wang. 2020. ConSTGAT: Contextual Spatial-Temporal Graph Attention Network for Travel Time Estimation at Baidu Maps. InProceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . 2697–2705

    work page 2020

  7. [7]

    Yushan Han, Hui Zhang, Huifang Li, Yi Jin, Congyan Lang, and Yidong Li. 2023. Collaborative perception in autonomous driving: Methods, datasets, and chal- lenges. IEEE Intelligent Transportation Systems Magazine (2023)

    work page 2023

  8. [8]

    Adam W Harley, Zhaoyuan Fang, Jie Li, Rares Ambrus, and Katerina Fragkiadaki

    work page

Show all 39 references
  1. [9]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition . 770–778

    2016

  2. [10]

    Minhyeok Heo, Jiwon Kim, and Sujung Kim. 2020. Hd map change detection with cross-domain deep metric learning. In 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) . IEEE, 10218–10224

    2020

  3. [11]

    Anthony Hu, Zak Murez, Nikhil Mohan, Sofía Dudas, Jeffrey Hawke, Vijay Badrinarayanan, Roberto Cipolla, and Alex Kendall. 2021. Fiery: Future instance prediction in bird’s-eye view from surround monocular cameras. In Proceedings of the IEEE/CVF International Conference on Comp...

    2021

  4. [12]

    Junjie Huang, Guan Huang, Zheng Zhu, Yun Ye, and Dalong Du. 2021. Bevdet: High-performance multi-camera 3d object detection in bird-eye-view. arXiv preprint arXiv:2112.11790 (2021)

    2021 arXiv

  5. [13]

    Jizhou Huang, Zhengjie Huang, Xiaomin Fang, Shikun Feng, Xuyi Chen, Jiaxiang Liu, Haitao Yuan, and Haifeng Wang. 2022. DuETA: Traffic Congestion Propaga- tion Pattern Modeling via Efficient Graph Learning for ETA Prediction at Baidu Maps. In Proceedings of the 31st ACM Interna...

    2022

  6. [14]

    Jizhou Huang, Haifeng Wang, Yibo Sun, Yunsheng Shi, Zhengjie Huang, An Zhuo, and Shikun Feng. 2022. ERNIE-GeoL: A Geography-and-Language Pre- trained Model and its Applications in Baidu Maps. InProceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Minin...

    2022

  7. [15]

    Kitae Kim, Soohyun Cho, and Woojin Chung. 2021. HD map update for au- tonomous driving with crowdsourced data. IEEE Robotics and Automation Letters 6, 2 (2021), 1895–1901

    2021

  8. [16]

    John Lambert and James Hays. 2022. Trust, but verify: Cross-modality fusion for hd map change detection. arXiv preprint arXiv:2212.07312 (2022)

    2022 arXiv

  9. [17]

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

    2022

  10. [18]

    Zhiqi Li, Wenhai Wang, Hongyang Li, Enze Xie, Chonghao Sima, Tong Lu, Yu Qiao, and Jifeng Dai. 2022. Bevformer: Learning bird’s-eye-view representa- tion from multi-camera images via spatiotemporal transformers. In European conference on computer vision . Springer, 1–18

    2022

  11. [19]

    Bencheng Liao, Shaoyu Chen, Xinggang Wang, Tianheng Cheng, Qian Zhang, Wenyu Liu, and Chang Huang. 2022. Maptr: Structured modeling and learning for online vectorized hd map construction. arXiv preprint arXiv:2208.14437 (2022)

    2022 arXiv

  12. [20]

    Bencheng Liao, Shaoyu Chen, Yunchi Zhang, Bo Jiang, Qian Zhang, Wenyu Liu, Chang Huang, and Xinggang Wang. 2023. Maptrv2: An end-to-end framework for online vectorized hd map construction.arXiv preprint arXiv:2308.05736 (2023)

    2023 arXiv

  13. [21]

    Yicheng Liu, Tianyuan Yuan, Yue Wang, Yilun Wang, and Hang Zhao. 2023. Vectormapnet: End-to-end vectorized hd map learning. InInternational Conference on Machine Learning. PMLR, 22352–22369

    2023

  14. [22]

    Ilya Loshchilov and Frank Hutter. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101 (2017)

    2017 arXiv

  15. [23]

    Bowen Pan, Jiankai Sun, Ho Yin Tiga Leung, Alex Andonian, and Bolei Zhou

  16. [24]

    David Pannen, Martin Liebner, and Wolfram Burgard. 2019. Hd map change detection with a boosted particle filter. In2019 International Conference on Robotics and Automation (ICRA). IEEE, 2561–2567

    2019

  17. [25]

    Lang Peng, Zhirong Chen, Zhangjie Fu, Pengpeng Liang, and Erkang Cheng

  18. [26]

    Limeng Qiao, Wenjie Ding, Xi Qiu, and Chi Zhang. 2023. End-to-End Vector- ized HD-Map Construction With Piecewise Bezier Curve. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 13218–13228

    2023

  19. [27]

    Andrew Tao, Karan Sapra, and Bryan Catanzaro. 2020. Hierarchical multi-scale attention for semantic segmentation. arXiv preprint arXiv:2005.10821 (2020)

    2020 arXiv

  20. [28]

    Ashley Varghese, Jayavardhana Gubbi, Akshaya Ramaswamy, and P Balamuralid- har. 2018. ChangeNet: A deep learning architecture for visual change detection. In Proceedings of the European conference on computer vision (ECCV) workshops . 0–0

    2018

  21. [29]

    Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, et al. 2020. Deep high-resolution representation learning for visual recognition. IEEE transactions on pattern analysis and machine intelligence 43, 10...

    2020

  22. [30]

    Deguo Xia, Jizhou Huang, Jianzhong Yang, Xiyan Liu, and Haifeng Wang. 2022. DuARUS: Automatic Geo-object Change Detection with Street-view Imagery for Updating Road Database at Baidu Maps. In Proceedings of the 31st ACM International Conference on Information & Knowledge Manag...

    2022

  23. [31]

    Deguo Xia, Xiyan Liu, Wei Zhang, Hui Zhao, Chengzhou Li, Weiming Zhang, Jizhou Huang, and Haifeng Wang. 2022. DuTraffic: Live traffic condition pre- diction with trajectory data and street views at Baidu maps. In Proceedings of the 31st ACM International Conference on Informat...

    2022

  24. [32]

    Deguo Xia, Weiming Zhang, Xiyan Liu, Wei Zhang, Chenting Gong, Jizhou Huang, Mengmeng Yang, and Diange Yang. 2024. DuMapNet: An End-to-End Vectorization System for City-Scale Lane-Level Map Generation. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and...

    2024

  25. [33]

    Congxi Xiao, Jingbo Zhou, Yixiong Xiao, Jizhou Huang, and Hui Xiong. 2024. ReFound: Crafting a Foundation Model for Urban Region Understanding upon Language and Visual Foundations. In Proceedings of the 30th ACM SIGKDD Con- ference on Knowledge Discovery and Data Mining . 3527–3538

    2024

  26. [34]

    Jianzhong Yang, Xiaoqing Ye, Bin Wu, Yanlei Gu, Ziyu Wang, Deguo Xia, and Jizhou Huang. 2022. DuARE: Automatic Road Extraction with Aerial Images and Trajectory Data at Baidu Maps. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 4321–4331

    2022

  27. [35]

    Zhixin Zhang, Yiyuan Zhang, Xiaohan Ding, Fusheng Jin, and Xiangyu Yue

  28. [39]

    arXiv preprint arXiv:2312.03341 (2023)

    Online Vectorized HD Map Construction using Geometry. arXiv preprint arXiv:2312.03341 (2023)

    2023 arXiv

  29. [2020]

    IEEE Robotics and Automation Letters 5, 3 (2020), 4867–4873

    Cross-view semantic segmentation for sensing surroundings. IEEE Robotics and Automation Letters 5, 3 (2020), 4867–4873

    2020

  30. [2022]

    arXiv e-prints (2022), arXiv–2206

    A simple baseline for bev perception without lidar. arXiv e-prints (2022), arXiv–2206

    2022

  31. [2023]

    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision

    BEVSegFormer: Bird’s Eye View Semantic Segmentation From Arbitrary Camera Rigs. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 5935–5943

Pith tools

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