Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Geometry-Guided Representations for Coherent Lane and Traffic Topology Reasoning in Driving Scenes

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read RelTopo shows that embedding geometric relations into lane perception, topology heads, and contrastive supervision jointly raises detection and topology accuracy, reporting new state-of-the-art scores on OpenLane-V2.

desk verdict A well-ablated multi-level relational model for lane topology, but the 'new SOTA' headline rests on baseline numbers whose metric version isn't established, so treat the absolute gains as provisional. read the letter →

arxiv 2506.13553 v4 pith:AUBEGK4E submitted 2025-06-16 cs.CV

Yueru Luo , Changqing Zhou , Yiming Yang , Erlong Li , Chao Zheng , Shuguang Cui , Zhen Li This is my paper · ORCID
classification cs.CV
keywords 3Dlanedetectiontopologyreasoninglane-to-laneconnectivitylane-to-traffic-elementassociationBéziercurverepresentationgeometry-biasedattentioncontrastiveInfoNCEsupervisionOpenLane-V2
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 road-topology reasoning fails when lane detection, lane-to-lane connectivity, and lane-to-traffic-element association are optimized separately, because none of those stages exploits the geometric structure of roads. Its proposed system, RelTopo, embeds relational cues at three levels: lane features are built with geometry-biased attention and curve-guided Bézier sampling; topology heads fuse geometry and cross-view features; and a contrastive loss pulls connected lane/traffic embeddings together. On the OpenLane-V2 benchmark, the paper reports consistent gains over prior methods, notably +5.3 in lane-to-lane topology and +4.9 in lane-to-traffic topology on subset A, with a new state-of-the-art overall score. If the comparison is fair, the result strengthens the case that perception and topology reasoning should be trained jointly rather than as a two-stage pipeline.

What carries the argument

The load-bearing object is the relation-aware lane decoder built on a cubic Bézier representation. Geometry-biased self-attention adds an MLP-embedded bias, computed from each lane pair's minimum end-to-start distance and orientation difference, to the attention logits; curve-guided cross-attention samples K=11 points along each Bézier curve and uses a single shared lane query to predict offsets and weights, so context is aggregated along the lane's shape rather than at sparse control points. This decoder feeds a geometry-enhanced L2L head, which broadcasts predecessor and successor query embeddings with sinusoidal positional encodings and adds a learned distance embedding of endpoint separation, and a cross-view L2T head, which projects 3D lane points into the front view, samples features by grid sampling, and forms pairwise lane-traffic embeddings. A contrastive InfoNCE loss with hard-negative mining regularizes the pairwise embeddings at supervision time.

What would settle it

Re-run the official OpenLane-V2 V2.1.0 evaluation on the released TopoLogic model and on RelTopo under identical training data and schedule; if the TOP_ll gap on subset A is not near +5.3 (and TOP_lt not near +4.9), the headline comparison is wrong even though the paper's internal ablations could still hold.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that relational modeling is not a post-processing step but a unifying inductive bias: injecting inter-lane geometry (distance and angle) directly into attention biases, sampling features along Bézier-curve lane representations, embedding endpoint distances into pairwise L2L logits, aligning BEV lane features to front-view traffic elements through projected geometry, and regularizing relational embeddings with a symmetric InfoNCE objective all cooperate. The paper demonstrates this with ablations: geometry-biased self-attention alone raises TOP_ll by +2.9, curve-guided cross-attention raises DET_l by +4.6, and the full stack reaches 48.9 OLS on subset A and 49.7 on subset B, outperforming tested baselines in every metric.

Load-bearing premise

The headline state-of-the-art claim assumes every baseline was evaluated under the same metric version (V2.1.0) and training protocol as RelTopo; the paper notes RoadPainter used old metrics and lists other baseline scores as sourced from the TopoLogic paper rather than recomputed.

Editorial extensions

If this is right

  • Gains appear in both perception and topology at once (DET_l +3.1, TOP_ll +5.3, TOP_lt +4.9 on subset A), supporting the paper's mutual-enhancement claim rather than a trade-off.
  • Topology predictions no longer depend on inference-time geometric post-processing; ensembling the paper's model with TopoLogic's geometric distance scoring actually lowers its TOP_ll by 0.5.
  • Lane-to-traffic association improves by aligning BEV lane features to front-view traffic features via projected 3D geometry, avoiding a second 2D lane decoder.
  • The relational modules are encoder-agnostic, improving OLS consistently with ResNet-50, ResNet-101, and Swin-Base backbones.

Reading between the lines

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

  • Editorial inference: the same three-level relational recipe could transfer to other pairwise map-reasoning tasks, such as lane-segment graphs or intersection topology, where distance/angle attention biases and contrastive relational supervision apply without the OpenLane-V2 task split.
  • Editorial inference: the size of the headline margin depends on metric-version alignment; if TopoLogic's published numbers come from an earlier evaluation, re-running both under V2.1.0 could narrow the +5.3 and +4.9 gains even though the internal ablations would still hold.
  • Editorial inference: since DET_t fluctuates across ablations while lane metrics rise, the design appears to move capacity between tasks; a gradient-balancing variant might stabilize traffic-element detection further.
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 / 6 minor

Summary. The paper proposes RelTopo, a unified framework for lane detection and road topology reasoning on the OpenLane-V2 benchmark. The method integrates relational modeling at three levels: perception (geometry-biased self-attention and curve-guided cross-attention in a Bezier-based lane decoder), reasoning (a geometry-enhanced L2L head and a cross-view L2T head), and supervision (a contrastive InfoNCE objective with hard-negative mining). The authors report state-of-the-art results on both subA and subB of OpenLane-V2, with gains of +3.1 DET_l, +5.3 TOP_ll, +4.9 TOP_lt, and +4.4 OLS on subA, and larger gains on subB. The manuscript also includes ablations, comparisons with alternative Bezier representations, and a backbone-generalization study.

Significance. If the reported results are reproducible and the comparisons are fair, the multi-level relational modeling idea is a solid contribution to driving-scene topology reasoning. The paper's internal ablations (Table 2) show monotonic improvements when adding each module, which supports the claim that relational perception, reasoning, and supervision complement each other. The appendix provides useful additional experiments, including a controlled comparison with TopoLogic's geometric distance topology (Table 3), alternative Bezier variants (Table 4), and consistent gains across ResNet-50, ResNet-101, and Swin-Base (Table 5). The authors state that code will be released, which would further strengthen reproducibility. However, the headline 'state-of-the-art' claim depends on the comparability of Table 1 baselines, which is not fully established.

major comments (4)
  1. [Table 1 and Section 4.3] The state-of-the-art claim rests on Table 1, but the comparability of the baseline numbers is not established. The table header states that all results use the V2.1.0 metric, yet the footnote discloses that RoadPainter‡ used the old metrics and that 'other results were sourced from the TopoLogic paper' without stating whether TopoLogic, TopoNet, MapTR, and VectorMapNet numbers were recomputed with the V2.1.0 evaluation code. Since TopoLogic (44.1 OLS) and TopoMLP (44.5 OLS) are the closest baselines on subA, a small metric-version shift in TOP_ll or TOP_lt could materially reduce the claimed +5.3 and +4.9 gains. Please clarify the exact evaluation protocol for each baseline and, if applicable, rerun the official V2.1.0 evaluation for all methods.
  2. [Table 1 (missing baseline)] Topo2D (Li et al., 2024), a method explicitly designed for L2T reasoning with 2D lane priors, is discussed in the related work but is absent from Table 1. Given that one of the paper's claimed strengths is +4.9/+6.0 TOP_lt gains, omitting the most directly competing L2T method makes the 'new state-of-the-art' statement incomplete even under a single metric version. Please include Topo2D in the comparison or justify its exclusion.
  3. [Table 2 and Sections 4.4, A.3] All experimental comparisons, including the central ablations in Table 2 and the GDT comparison in Appendix Table 3, are reported without error bars or multiple seeds. Differences of 0.6-1.7 points in TOP_ll and TOP_lt could plausibly arise from training variance, especially given the DET_t fluctuations the authors themselves analyze in Appendix A.3.5. Please report the mean and standard deviation over at least three seeds for the main ablations, or explicitly justify the stability of single-run results.
  4. [Section 4.4 and Appendix A.2.1] The ablation baseline (#1) is the authors' own reimplementation of a deformable-DETR decoder with lightweight MLP heads, said to follow Wu et al. (2023). The baseline DET_l is 27.7, whereas Table 1 lists TopoMLP's DET_l as 28.5. This discrepancy raises the question of whether the baseline faithfully reproduces TopoMLP or differs in implementation details. Please state whether the ablation baseline matches the published TopoMLP numbers, and discuss how any implementation gap affects the attribution of gains to the proposed modules.
minor comments (6)
  1. [Abstract and Title] The model is called 'CoPo' in the abstract but 'RelTopo' in the title, introduction, and the rest of the paper. Please unify the name.
  2. [Section 4.3 and Table 1] The text repeatedly says 'OSL' (e.g., 'RelTopo achieves an OSL of 48.9') but the metric is abbreviated OLS in Eq. (10) and Table 1. Please correct the typo.
  3. [Figure 2 caption] The caption text is garbled: 'associations between lanes , and among lanes and associations between lanes and traffic elements (L2T)' repeats phrases and the inline symbols are missing. Please rewrite the caption to clearly describe the pipeline, including the definitions of Q_lane, Q_te, P_lane, and P_te.
  4. [Equation (9)] The multi-positive extension of InfoNCE is written as log[1 + sum_{v+} sum_{v-} exp(v- - v+)]. This treats each positive-negative pair independently rather than the standard multi-positive softmax. Please clarify the derivation and cite the exact source; if this is intentional, explain how it relates to the usual multi-positive contrastive loss.
  5. [Throughout] The name 'Topologic' is used inconsistently alongside 'TopoLogic' (e.g., in Sections 3.2.1 and 3.3.1). Please use the official spelling 'TopoLogic' consistently.
  6. [Appendix Table 4] The table title contains a typo: 'Comparsion' should be 'Comparison'.

Circularity Check

0 steps flagged · score 0.0 of 10

RelTopo's derivation is self-contained: relational cues are model inputs and external benchmark labels are the targets, so no prediction reduces by construction to its inputs.

full rationale

No circular step meets the evidence bar. The perception-level geometry biases (Dist, Angle) and the L2L distance embedding are features computed from predicted lane geometry; the L2L/L2T targets are externally annotated connectivity labels, so the mapping from features to labels is learned under a supervised loss and tested on held-out data. The contrastive InfoNCE term uses the same labels only as training supervision; it does not define the reported metrics. Table 1 is an external benchmark comparison; the footnote disclosure that some baseline numbers come from prior papers, and that RoadPainter used older metrics, is a comparability and correctness concern, not a derivation that reduces a prediction to its input. The only self-citation (Luo et al. 2023) appears in related work and is not load-bearing; no uniqueness theorem or ansatz is imported from the authors' prior work. The ablations use the authors' own reimplementation as the baseline, but they are accompanied by external comparisons and do not by themselves force the headline result. Therefore the paper is not circular; any dispute about the state-of-the-art claim should be framed as an external reproducibility and metric-version question, not as a circularity defect.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard ML assumptions (benchmark validity, representation fidelity) plus a few method-specific design choices. The most fragile are the fairness of the baseline comparison and the correctness of the contrastive loss formula, both of which are unverified without code or a corrected derivation.

free parameters (6)
  • lambda_7 (topology focal loss weight) = 5.0
    Hand-chosen weight balancing the topology focal loss against the contrastive loss in the total relation loss.
  • lambda_8 (contrastive loss weight) = 0.1
    Hand-chosen weight for the InfoNCE regularization term.
  • K (sampled points per Bezier curve) = 11
    Number of reference points sampled along each cubic Bezier curve in curve-guided cross-attention.
  • number of negative samples in contrastive loss = 3
    Each contrastive pair includes one positive and three negatives, selected via hard-negative mining.
  • number of decoder layers = 6
    Six decoder layers for both lane and traffic element decoders, following prior work.
  • number of lane queries = 300
    Number of learnable lane queries in the DETR decoder, following Wu et al. (2023).
assumptions (5)
  • domain assumption OpenLane-V2 ground truth and the V2.1.0 evaluation metric are correct and comparable across baselines.
    The SOTA claim relies on the benchmark annotations and metric version being applied consistently to all methods. The paper only partially verifies this.
  • domain assumption A single cubic Bezier curve with four control points adequately represents each lane.
    The lane detector and curve-guided cross-attention depend on this representation; the paper compares with BeMapNet and BezierFormer but not all alternatives.
  • domain assumption OpenLane-V2 camera intrinsics and extrinsics are accurate for projecting 3D lane points to the front-view plane.
    The cross-view L2T head's feature alignment uses this projection; inaccurate calibration would break the alignment.
  • ad hoc to paper The multi-positive InfoNCE extension in Eq. (9) is correct as written.
    The formula appears inconsistent with standard multi-positive InfoNCE, and the cited reference (Wang et al. 2023) is unrelated to contrastive learning.
  • ad hoc to paper Hard-negative mining on predicted logits during training stabilizes learning and does not bias the model.
    Selecting top-n negatives by model logits is a design choice that could reinforce model errors; its effects are not analyzed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Geometry-Guided Representations for Coherent Lane and Traffic Topology Reasoning in Driving Scenes." pith.science (2026). https://pith.science/paper/AUBEGK4E

@misc{pith2026250613553,
  author       = {Pith},
  title        = {Pith review of: Geometry-Guided Representations for Coherent Lane and Traffic Topology Reasoning in Driving Scenes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AUBEGK4E}},
  note         = {Machine review of arXiv:2506.13553}
}
abstract

Road topology reasoning is fundamental for autonomous driving, requiring both accurate perception of road elements and understanding of their complex connectivity, including lane connectivity (Lane-to-Lane, L2L) and traffic regulation (Lane-to-Traffic signs, L2T). However, existing methods typically treat perception and topology reasoning as fragmented tasks, ignoring their potential for mutual enhancement. Crucially, while topology is inherently relational, prior works often overlook geometric relationships during feature extraction, relying instead on brittle post-processing or coordinate-based heuristics applied only at inference time. To bridge this gap, we propose CoPo (Coherent Perception and toPology), a unified framework that integrates geometry-guided relational modeling across three levels: 1) Perception-level: We introduce a relation-aware lane detector that utilizes geometry-biased self-attention and curve-guided cross-attention to enrich lane representations with structural priors; 2) Reasoning-level: We design relation-enhanced topology heads, including a geometry-enhanced L2L head and a cross-view L2T head, which effectively align features to infer connectivity; and 3) Supervision-level: We implement a contrastive InfoNCE strategy to regularize relational embeddings, pulling connected pairs closer in the latent space. This coherent multi-level design enables end-to-end joint optimization of perception and reasoning. Extensive experiments on OpenLane-V2 demonstrate that CoPo significantly outperforms existing methods, achieving gains of {\textbf{+3.1}} in DET$_l$, {\textbf{+5.3}} in TOP$_{ll}$, {\textbf{+4.9}} in TOP$_{lt}$, and {\textbf{+4.4}} overall in OLS, setting a new state-of-the-art.

Figures

Figures reproduced from arXiv: 2506.13553 by the authors.

Figure 1
Figure 1. Performance and visualization comparison between previous methods and ours. The top-left shows multi-view input images. Our approach, RelTopo, integrates relational modeling across multiple levels, strengthening both perception and topology reasoning. As shown in the bottom-left quantitative results on OpenLane-V2, RelTopo significantly outperforms prior methods across all metrics. On the right, qualitative comparis… view at source ↗
Figure 2
Figure 2. The overall framework of RelTopo , which embeds relational modeling across three levels. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Illustration of our Bezier lane de- ´ coder layer, featuring our geometry-biased SA and curve-guided CA . By enhancing geometrically proximal lanes, it allows the model to allocate greater attention to spatially rel￾evant lanes through the interleaved attention process, leading to a more robust understanding of lane topology (detailed in Sec. 3.3.1). Our method differs from Hou et al. (2025), which encodes progressi… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Comparative visual results on OpenLane-V2 [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Qualitative results comparison. The 1st row presents the multi-view input images, the 2 nd row shows the predicted lane centerline results, and the 3rd row illustrates the predicted L2L topology results, with light blue arrowlines indicating directed connectivity betwe…
Figure 6
Figure 6. Figure 6: More qualitative result comparison. With our proposed designs, RelTopo achieves more [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. HGeo-TopoMap: Boosting Topological Mapping with Hierarchical Geometric Priors

    cs.CV 2026-07 conditional novelty 6.0 of 10

    HGeo-TopoMap injects explicit road-structure maps and implicit geometric relations into a DETR-style detector, improving top-down centerline mapping and topology reasoning on OpenLane-V2.

Reference graph

Works this paper leans on

22 extracted references · 5 canonical work pages · cited by 1 Pith paper

  1. [6]

    Leveraging our proposed components, RelTopo demonstrates superior perception of lane centerlines compared to previous methods, as shown in the second row of the figures. Moreover, RelTopo achieves more accurate lane-to-lane (L2L) 16 Under review Multi-view InputsLane CenterlinesL2L TopoFront-view L2T Topo GTTopoMLPOursGTTopoMLPOurs Figure 5: Qualitative r...

    work page 2023

  2. [7]

    Reconstruct from top view: A 3d lane detection approach based on geometry structure prior

    Chenguang Li, Jia Shi, Ya Wang, and Guangliang Cheng. Reconstruct from top view: A 3d lane detection approach based on geometry structure prior. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4370–4379, 2022a. Han Li, Zehao Huang, Zitian Wang, Wenge Rong, Naiyan Wang, and Si Liu. Enhancing 3d lane detection and to...

    arXiv

  3. [8]

    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. In2022 International Conference on Robotics and Automation (ICRA), pp. 4628–4634. IEEE, 2022b. Tianyu Li, Li Chen, Huijie Wang, Yang Li, Jiazhi Yang, Xiangwei Geng, Shengyin Jiang, Yuting Wang, Hang Xu, Chunjing Xu, et al. Graph-based topology rea...

    arXiv

  4. [9]

    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. arXiv preprint arXiv:2308.05736,

    arXiv

  5. [10]

    Focal loss for dense object detection.arXiv preprint arXiv:1708.02002,

    T Lin. Focal loss for dense object detection.arXiv preprint arXiv:1708.02002,

    arXiv

  6. [11]

    Petr: Position embedding transformation for multi-view 3d object detection

    Yingfei Liu, Tiancai Wang, Xiangyu Zhang, and Jian Sun. Petr: Position embedding transformation for multi-view 3d object detection. InEuropean Conference on Computer Vision, pp. 531–548. Springer, 2022b. Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shift...

    arXiv

  7. [12]

    Roadpainter: Points are ideal navigators for topology transformer.arXiv preprint arXiv:2407.15349,

    Zhongxing Ma, Shuang Liang, Yongkun Wen, Weixin Lu, and Guowei Wan. Roadpainter: Points are ideal navigators for topology transformer.arXiv preprint arXiv:2407.15349,

    arXiv

  8. [13]

    Representation learning with contrastive predictive coding.arXiv preprint arXiv:1807.03748,

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Representation learning with contrastive predictive coding.arXiv preprint arXiv:1807.03748,

    arXiv

Show all 22 references
  1. [14]

    Instagram: Instance-level graph modeling for vectorized hd map learning.arXiv preprint arXiv:2301.04470,

    Juyeb Shin, Francois Rameau, Hyeonjun Jeong, and Dongsuk Kum. Instagram: Instance-level graph modeling for vectorized hd map learning.arXiv preprint arXiv:2301.04470,

    arXiv

  2. [15]

    Making large language models better reasoners with alignment.arXiv preprint arXiv:2309.02144,

    Peiyi Wang, Lei Li, Liang Chen, Feifan Song, Binghuai Lin, Yunbo Cao, Tianyu Liu, and Zhifang Sui. Making large language models better reasoners with alignment.arXiv preprint arXiv:2309.02144,

    arXiv

  3. [16]

    Argoverse 2: Next generation datasets for self-driving perception and forecasting.arXiv preprint arXiv:2301.00493,

    Benjamin Wilson, William Qi, Tanmay Agarwal, John Lambert, Jagjeet Singh, Siddhesh Khandelwal, Bowen Pan, Ratnesh Kumar, Andrew Hartnett, Jhony Kaesemodel Pontes, et al. Argoverse 2: Next generation datasets for self-driving perception and forecasting.arXiv preprint arXiv:2301.00493,

    arXiv

  4. [17]

    Topomlp: An simple yet strong pipeline for driving topology reasoning.arXiv preprint arXiv:2310.06753,

    Dongming Wu, Jiahao Chang, Fan Jia, Yingfei Liu, Tiancai Wang, and Jianbing Shen. Topomlp: An simple yet strong pipeline for driving topology reasoning.arXiv preprint arXiv:2310.06753,

    arXiv

  5. [18]

    Online map vectorization for autonomous driving: A rasterization perspective.Advances in Neural Information Processing Systems, 36:31865–31877, 2023a

    12 Under review Gongjie Zhang, Jiahao Lin, Shuang Wu, Zhipeng Luo, Yang Xue, Shijian Lu, Zuoguan Wang, et al. Online map vectorization for autonomous driving: A rasterization perspective.Advances in Neural Information Processing Systems, 36:31865–31877, 2023a. Zhixin Zhang, Yi...

    arXiv

  6. [19]

    Deformable detr: Deformable transformers for end-to-end object detection.arXiv preprint arXiv:2010.04159,

    Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, and Jifeng Dai. Deformable detr: Deformable transformers for end-to-end object detection.arXiv preprint arXiv:2010.04159,

    2010 arXiv

  7. [20]

    13 Under review A APPENDIX A.1 LLM USAGE In preparing this manuscript, we used a large language model (LLM) to help polish phrasing and grammar. A.2 IMPLEMENTATIONDETAILS A.2.1 OVERALLOPTIMIZATIONOBJECTIVE As discussed in the main paper, our training objective comprises detect...

    2023

  8. [21]

    Topologic – GDT L2L

    The corresponding weights areλ4 = 1.5 for classification,λ5 = 0.05 for L1 loss, andλ6 = 0.02 for the chamfer distance loss. For topology reasoning, we adopt a focal loss for classification, as in prior work, and introduce a contrastive lossLcon to further enhance relational le...

    2024

  9. [2019]

    Gen-lanenet: A generalized and scalable approach for 3d lane detection

    Yuliang Guo, Guang Chen, Peitao Zhao, Weide Zhang, Jinghao Miao, Jingao Wang, and Tae Eun Choe. Gen-lanenet: A generalized and scalable approach for 3d lane detection. InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXI...

    2020

  10. [2020]

    Topo- logic: An interpretable pipeline for lane topology reasoning on driving scenes.arXiv preprint arXiv:2405.14747,

    Yanping Fu, Wenbin Liao, Xinyuan Liu, Yike Ma, Feng Dai, Yucheng Zhang, et al. Topo- logic: An interpretable pipeline for lane topology reasoning on driving scenes.arXiv preprint arXiv:2405.14747,

    arXiv

  11. [2022]

    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. InComputer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part I 16, pp. 213–...

    2020

  12. [2023]

    Topobda: Towards bezier deformable attention for road topology understanding.arXiv preprint arXiv:2412.18951,

    Muhammet Esat Kalfaoglu, Halil Ibrahim Ozturk, Ozsel Kilinc, and Alptekin Temizel. Topobda: Towards bezier deformable attention for road topology understanding.arXiv preprint arXiv:2412.18951,

  13. [2024]

    3d-lanenet+: Anchor free lane detection using a semi-local representation.arXiv preprint arXiv:2011.01535,

    Netalee Efrat, Max Bluvstein, Shaul Oron, Dan Levi, Noa Garnett, and Bat El Shlomo. 3d-lanenet+: Anchor free lane detection using a semi-local representation.arXiv preprint arXiv:2011.01535,

    2011 arXiv

  14. [2025]

    Admap: Anti-disturbance framework for reconstructing online vectorized hd map.arXiv preprint arXiv:2401.13172,

    Haotian Hu, Fanyi Wang, Yaonong Wang, Laifeng Hu, Jingwei Xu, and Zhiwang Zhang. Admap: Anti-disturbance framework for reconstructing online vectorized hd map.arXiv preprint arXiv:2401.13172,

    arXiv

Pith tools

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