Pith. sign in

REVIEW 4 major objections 7 minor 60 references

HV-BEV: Decoupling Horizontal and Vertical Feature Sampling for Multi-View 3D Object Detection

T0 review · 4 major / 7 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that decoupling bird's-eye-view feature sampling into adaptive height-aware vertical sampling and dynamic horizontal cross-view neighbor aggregation improves multi-view 3D object detection, raising base-validation mAP…

desk verdict A solid BEVFormer upgrade with a genuinely new height-aware sampling mechanism; the main gain is real but needs a compute-matched control to prove it comes from the decoupled design. read the letter →

arxiv 2412.18884 v3 pith:43G5X5TA submitted 2024-12-25 cs.CV

Di Wu , Feng Yang , Benlian Xu , Pan Liao , Wenhui Zhao , Dingwen Zhang This is my paper · ORCID
classification cs.CV
keywords multi-view3Dobjectdetectionbird's-eyeviewadaptiveheight-awaresamplingcross-viewfeatureaggregationdeformableattentiontemporalfusionnuScenesbenchmarkBEVgridqueries
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 proposes a dense bird's-eye-view (BEV) multi-camera 3D object detector that separates feature sampling into two directions. On the horizontal plane, each 3D reference point learns a small set of neighboring points across grid cells and camera views, so an object spanning many BEV cells or several images is aggregated as one structure. In the vertical direction, instead of sampling reference points uniformly over a fixed height range, the model predicts a height distribution per grid cell from current and temporally aligned historical BEV features and samples heights near where objects of that scene actually sit. The paper reports that these two changes raise validation mAP from 41.6% to 43.9% and NDS from 51.7% to 53.3% in the base configuration over the grid-query baseline, and 50.5% mAP and 59.8% NDS on the nuScenes test set.

What carries the argument

The load-bearing mechanism is the decoupled sampling scheme applied inside each BEV encoder layer. A vertical adaptive height-aware sampling module takes the current BEV queries and the aligned historical BEV features, feeds both through a shared MLP into $D=8$ discrete height bins, combines the two distributions with a dual-branch cross-attention fusion, and selects the top $N_{ref}=4$ heights as the z-coordinates of each query's 3D reference points. A dynamic horizontal cross-view feature aggregation module then learns 2D offsets on the ground-aligned plane for each reference point, forming a four-neighbor graph whose nodes are projected into the images and aggregated by learned softmax edge weights. The two modules replace uniform global-height sampling and isolated single-point cross-attention, so the model samples where height information is concentrated and shares features across the spatial extent of an object.

What would settle it

Train the vertical height-aware module while corrupting the predicted height distributions at inference, for example replacing the fused distribution with a uniform one or with random heights while keeping the horizontal module intact; if mAP and NDS do not drop below the intact model, then the adaptive heights themselves carry no object-height information and the module's reported gain must come from something else.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the two quantities BEV-grid query sampling usually collapses—where along the height axis an object sits, and which neighboring grid cells belong to the same object—can be recovered explicitly and separately, and that doing so improves multi-view 3D detection. The vertical module fuses a predicted current height distribution with a historical distribution warped by ego-motion, supervises the fused distribution with annotated center heights of 3D boxes, and uses top-k sampling to place reference points at object-relevant heights. The horizontal module then treats each reference point as the center of a dynamically learned neighborhood on its height plane, projects all neighboring points into the camera views, and combines their bilinearly sampled features with learned edge weights. The paper reports that both modules together beat the baseline consistently across three configurations on nuScenes, that the horizontal module contributes the larger share of the gain in ablations, and that the same BEV features improve 3D semantic occupancy prediction when the task head is replaced.

Load-bearing premise

The whole vertical module rests on the assumption that the flattened 2D bird's-eye-view features already contain enough hidden information about object heights for the shared MLP to predict where to sample; if that height signal is not actually present in the BEV features, the adaptive sampling just moves reference points around without adding information.

Editorial extensions

If this is right

  • If the gains reproduce, dense BEV-grid perception can keep a single shared representation for detection and other tasks while still exploiting object height, because the added modules live inside the feature encoder rather than in a task-specific sparse query head.
  • The base configuration's 2.3-point mAP and 1.6-point NDS improvements imply that structured horizontal aggregation, not just more reference points, is what recovers objects spanning multiple BEV cells or views, since increasing uniform samples from 4 to 8 gives only 0.4 mAP in the paper's control experiment.
  • The occupancy experiments imply the same height-aware BEV features are not detection-specific: replacing the detection head with a voxel segmentation head raised scene-completion IoU by 1.28 points and semantic mIoU by 3.22 points over the baseline.
  • The Lyft results imply the approach is not overfit to nuScenes camera geometry, at least under the evaluation protocol used there, beating the camera and LiDAR baselines listed in the paper on that validation split.

Reading between the lines

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

  • Inference: the ablation pattern—VHA alone gains 0.5 mAP while DHCA alone gains 2.2 mAP—suggests the headline improvement is mostly driven by horizontal neighborhood aggregation, and the vertical module's contribution may be to make those neighbors more informative rather than to add large independent accuracy; an experiment running DHCA with fixed uniform heights would separate these two readings.
  • Inference: the paper supervises heights only from annotated 3D box centers; if the same module were trained with dense or voxel-level height labels, the predicted height distribution might also serve as a lightweight 3D geometric prior for downstream planning, not just a sampling guide.
  • Inference: the historical-height fusion assumes the ego vehicle only rotates around the vertical axis and translates horizontally; on roads with significant pitch or roll, the 2D warping could misalign the historical height distribution, so a 6-DoF alignment or learning-based correction is a natural extension the paper does not test.
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 / 7 minor

Summary. The paper proposes HV-BEV, a multi-view 3D object detector built on BEVFormer that decouples feature sampling into a horizontal and a vertical component. The vertical adaptive height-aware (VHA) module predicts a discrete height distribution from current and historical BEV features, fuses the two predictions with a dual-branch cross-attention, and uses a top-k selection to choose reference-point heights, supervised by Gaussian-smoothed ground-truth center heights. The dynamic horizontal cross-view aggregation (DHCA) module learns 2D offsets in the BEV plane for each 3D reference point, projects the resulting neighboring points into the camera views, and aggregates their bilinearly sampled features with learned edge weights. Experiments on nuScenes validation and test sets and on Lyft report consistent improvements over the BEVFormer baseline, with the base configuration improving mAP from 41.6% to 43.9% and NDS from 51.7% to 53.3% on nuScenes validation, and the largest configuration reaching 50.5% mAP and 59.8% NDS on the nuScenes test set. The paper also reports a transfer of the modules to 3D occupancy prediction.

Significance. If the reported gains are attributable to the proposed decoupled horizontal and vertical sampling, the work is a useful contribution to dense BEV perception: it preserves the multi-task extendability of BEV grid queries while adding object-aware height sampling and cross-view feature aggregation. The paper ships code, evaluates on two datasets, and provides ablations for both modules, which is a strength. The main significance caveat is that the ablation evidence does not yet establish that the gains come from the decoupled design rather than from a larger number of sampled image features or from increased model capacity. The VHA module in isolation adds only 0.5 mAP and 0.3 NDS, and the DHCA control in Table IX does not equalize sampling counts or compute. Consequently, the central attribution claim is plausible but not fully supported.

major comments (4)
  1. [Table IX, Sec. IV-E] The control experiment in Table IX varies only the number of uniformly sampled reference points Nref in the vanilla BEVFormer (4, 8, 16). It does not equalize the number of sampled feature locations per query with HV-BEV, nor does it equalize compute. In HV-BEV, each of the Nref reference points first computes x_p with deformable attention and two sampling points per view (Sec. IV-B), and then samples M=4 neighboring points with bilinear interpolation in every hit view (Eq. 14). This multiplies the number of image-feature locations per query by roughly a factor of 1+M/n_deform relative to a baseline that uses n_deform deformable points per reference point. The results in Table IX therefore cannot rule out the explanation that the DHCA gain comes from sampling more image features or from the extra learnable offsets, rather than from the horizontal graph structure itself. I ask for an additional baseline that matches the total number of sampled locations and approximately the same FLOPs/FPS, for example a BEVFormer variant with more deformable sampling points per reference point or with an equivalent number of additional bilinear samples arranged in a non-graph (e.g., random or uniform) configuration. Without such a control, the central claim that the decoupled horizontal design is responsible for the improvement is not established.
  2. [Table VI, Sec. IV-E] The VHA module alone yields only +0.5 mAP and +0.3 NDS over the small-configuration baseline, and the paper reports a single run for each configuration. Given typical run-to-run variance in end-to-end trained detection models with this training setup, this difference is not demonstrably significant. The paper should provide at least three seeds with mean and standard deviation for the main ablation rows, or otherwise justify that the VHA-only gain is outside noise. This matters because the paper's central narrative attributes part of the improvement to adaptive height sampling.
  3. [Sec. III-B, Sec. III-C] The height-aware module selects the height coordinates of the Nref reference points by applying a top-k operation to the fused height distribution H_fus (Sec. III-B: 'we apply a top-k function to sample Nref discrete height values'). Top-k selection is not differentiable, and the manuscript does not specify a straight-through estimator, Gumbel-top-k, or a continuous relaxation. As written, gradients from the detection loss cannot flow through the selected height indices into H_fus; only the auxiliary height loss L_hgt (Eq. 8) can supervise H_fus. This creates an inconsistency with the claim that the model is trained end-to-end and leaves unclear how the height-sampling pathway is optimized for the detection objective. Please state the exact sampling/differentiation procedure and, if the top-k is used as a non-differentiable selector, discuss the implications and provide evidence that the height distribution is nevertheless learned effectively.
  4. [Eq. (3)-(4), Sec. III-C] The temporal warp of the historical height distribution uses a 2D rotation matrix and translation vector around the vertical z axis only. This ignores vehicle pitch and roll, which can be non-negligible on slopes and could misalign the historical height distribution with the current grid. The paper does not discuss this assumption or provide evidence that the 2D warp is sufficient. Since the baseline BEVFormer also operates in a flattened BEV plane, this may be an inherited limitation rather than a new error, but it should be stated explicitly and, if possible, quantified on nuScenes (e.g., by comparing against a warp that uses the full 6-DoF ego pose).
minor comments (7)
  1. [Sec. III-C, Eq. (6)] The conditional definition of H_gt_{p,m} for the two cases I_p=1 and I_p=0 is not typeset clearly in Eq. (6); please separate the cases explicitly so the uniform distribution for background cells is unambiguous.
  2. [Sec. IV-B, Eq. (16)] The loss weights lambda_1, lambda_2, and lambda_3 in Eq. (16) are introduced but their numerical values are never reported. Please provide the values used in all experiments.
  3. [Sec. III-C] After the top-k operation selects height bins, the manuscript does not state how the discrete bins are converted to continuous height values for the 3D reference points (e.g., bin centers or interpolation). Please specify this conversion.
  4. [Sec. IV-B] The number of deformable attention sampling points used by the baseline BEVFormer spatial cross-attention is not stated, which is needed to compute the sample-count ratio discussed in the major comments. Please report this value for both baseline and HV-BEV.
  5. [Sec. V] The conclusion states that the introduced loss provides 'depth supervision,' but the loss in Eq. (8) supervises height distributions, not depth. This appears to be a typo and should be corrected.
  6. [Sec. IV-C] In the comparison with state-of-the-art, the phrase 'surpassing one of the latest methods' is supported by Table II, but the image sizes differ slightly between HV-BEV small (1280x720) and BEVFormer-small-QAF2D (1280x736). Please note whether this difference affects the comparison.
  7. [Throughout] There are several typographical and formatting issues, including 'odject' in Sec. IV-C, the rendered 'V oVnet' in Tables and text, and the citation for the Lyft dataset missing a full bibliographic entry. These should be cleaned up.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: HV-BEV is an empirical architecture study validated on an external benchmark, with no self-referential reduction of the claimed results to their inputs.

full rationale

The paper's central claim is an empirical improvement over BEVFormer on nuScenes, supported by held-out validation and test metrics. The height-aware module is supervised by annotated 3D box centers through Eqs. 6-8, and the detection metrics are measured on the validation and testing splits; no target metric is used as a training signal or fitted parameter. The DHCA neighbor offsets and edge weights are learned functions of BEV queries (Eqs. 9-10, 15), not fitted to the reported mAP or NDS. The baseline comparison uses official BEVFormer configurations, and the ablation tables decompose the gains into the VHA and DHCA components. The skeptic concern that Table IX does not equalize the number of sampled image features or compute between the baseline and HV-BEV is a legitimate threat to the attribution of the gain to the decoupled design, but it is a fairness or validity issue rather than circularity: it does not make any predicted quantity equivalent to an input by construction. No load-bearing self-citation chain appears; references to BEVFormer, BEV-SAN, HeightFormer, and similar works are external prior art used as baselines or motivation. Therefore no circular step can be exhibited, and the appropriate circularity score is 0.

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

All free parameters are standard architecture hyperparameters, tuned either by hand or via ablation on the validation set. The paper introduces no new physical entities. The main assumptions are dataset and calibration standard, with two ad hoc assumptions about the trainability and information content of the height prediction that are load-bearing and not fully justified.

free parameters (5)
  • Number of height bins (D) = 8
    Ablated in Table VII; chosen as best on nuScenes validation. Controls vertical resolution of the height distribution and affects the density of reference point sampling.
  • Number of neighboring points (M) = 4
    Ablated in Table VIII; chosen as a trade-off between accuracy (4 and 8 are similar) and memory/FPS. Higher values degrade accuracy via over-smoothing.
  • Number of reference points (Nref) = 4
    Inherited from BEVFormer; Table IX tests Nref 4/8/16 on the baseline. HV-BEV uses Nref=4 with M=4 neighbors.
  • Gaussian kernel sigma = 1
    Set in Sec IV-B for building the ground-truth height supervision; not ablated.
  • Loss weights lambda1, lambda2, lambda3 = not reported
    Used in Eq. (16) to balance detection and height losses; values are not given in the paper, making precise replication harder.
assumptions (6)
  • standard math Deformable attention is a valid differentiable sampling mechanism
    Used throughout the BEVFormer baseline and the new modules (Eqs. 1,2,5,11); the formulation from [19] is assumed as background.
  • domain assumption Camera intrinsic parameters are known and accurate
    The projection P(p3d, i) in Eq. (13) uses the intrinsic matrix; calibration errors would degrade sampling. This is standard for the nuScenes dataset.
  • domain assumption Ego-motion between consecutive frames is known and can be represented by a 2D rotation and translation
    The mapping M in Eq. (3)-(4) warps historical height distributions to the current frame using R (2x2) and T (2x1). This ignores pitch/roll of the vehicle.
  • domain assumption Ground-truth 3D bounding box center heights are available for all objects in the training set
    The height distribution supervision (Eqs. 6-8) uses z_gt from annotated boxes; the method cannot be trained on unlabeled data.
  • ad hoc to paper The flattened BEV features contain enough information to predict object height distribution
    Sec III-C predicts Hcur from BEV queries Q' which have no explicit z dimension; whether they encode height is an assumption not directly verified in the paper.
  • ad hoc to paper Top-k selection of Nref heights can be integrated into end-to-end training
    Sec III-C says top-k is used to sample discrete height values, but no differentiable relaxation or straight-through estimator is described. The paper silently assumes this is trainable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HV-BEV: Decoupling Horizontal and Vertical Feature Sampling for Multi-View 3D Object Detection." pith.science (2026). https://pith.science/paper/43G5X5TA

@misc{pith2026241218884,
  author       = {Pith},
  title        = {Pith review of: HV-BEV: Decoupling Horizontal and Vertical Feature Sampling for Multi-View 3D Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/43G5X5TA}},
  note         = {Machine review of arXiv:2412.18884}
}
read the original abstract

The application of vision-based multi-view environmental perception system has been increasingly recognized in autonomous driving technology, especially the BEV-based models. Current state-of-the-art solutions primarily encode image features from each camera view into the BEV space through explicit or implicit depth prediction. However, these methods often overlook the structured correlations among different parts of objects in 3D space and the fact that different categories of objects often occupy distinct local height ranges. For example, trucks appear at higher elevations, whereas traffic cones are near the ground. In this work, we propose a novel approach that decouples feature sampling in the \textbf{BEV} grid queries paradigm into \textbf{H}orizontal feature aggregation and \textbf{V}ertical adaptive height-aware reference point sampling (HV-BEV), aiming to improve both the aggregation of objects' complete information and awareness of diverse objects' height distribution. Specifically, a set of relevant neighboring points is dynamically constructed for each 3D reference point on the ground-aligned horizontal plane, enhancing the association of the same instance across different BEV grids, especially when the instance spans multiple image views around the vehicle. Additionally, instead of relying on uniform sampling within a fixed height range, we introduce a height-aware module that incorporates historical information, enabling the reference points to adaptively focus on the varying heights at which objects appear in different scenes. Extensive experiments validate the effectiveness of our proposed method, demonstrating its superior performance over the baseline across the nuScenes dataset. Moreover, our best-performing model achieves a remarkable 50.5\% mAP and 59.8\% NDS on the nuScenes testing set. The code is available at https://github.com/Uddd821/HV-BEV.

Figures

Figures reproduced from arXiv: 2412.18884 by the authors.

Figure 1
Figure 1. The multi-view feature sampling comparison between (a) BEVFormer [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall architecture of HV-BEV. Aligning with BEVFormer, we begin by using an image encoder to extract multi-view image features. Within the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The dual-branch cross-attention (DCA) fusion strategy between the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: The process of the dynamic horizontal cross-view feature aggregation [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Visualization results of HV-BEV and the baseline on nuScenes validation set. Different colors of 3D bounding boxes denote different object classes. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Robust visualization results on Lyft validation set. Different colors of [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: A qualitative comparison of 3D semantic occupancy prediction results between SurroundOcc and SurroundOcc-DHCA on two samples from the [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 31 canonical work pages

  1. [1]

    3d object detection for autonomous driving: A comprehensive survey,

    J. Mao, S. Shi, X. Wang, and H. Li, “3d object detection for autonomous driving: A comprehensive survey,” International Journal of Computer Vision, vol. 131, no. 8, pp. 1909–1963, 2023

    1909

  2. [2]

    V oxelnet: End-to-end learning for point cloud based 3d object detection,

    Y . Zhou and O. Tuzel, “V oxelnet: End-to-end learning for point cloud based 3d object detection,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 4490–4499

    2018

  3. [3]

    Multi-modal 3d object detection by box matching,

    Z. Liu, X. Ye, Z. Zou, X. He, X. Tan, E. Ding, J. Wang, and X. Bai, “Multi-modal 3d object detection by box matching,” IEEE Transactions on Intelligent Transportation Systems , 2024

    work page 2024

  4. [4]

    Scnet3d: Rethinking the feature extraction process of pillar-based 3d object detection,

    J. Li, Z. Wang, D. Gong, and C. Wang, “Scnet3d: Rethinking the feature extraction process of pillar-based 3d object detection,” IEEE Transactions on Intelligent Transportation Systems , 2024

    work page 2024

  5. [5]

    Vision-centric bev perception: A survey,

    Y . Ma, T. Wang, X. Bai, H. Yang, Y . Hou, Y . Wang, Y . Qiao, R. Yang, and X. Zhu, “Vision-centric bev perception: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

    work page 2024

  6. [6]

    Occformer: Dual-path transformer for vision-based 3d semantic occupancy prediction,

    Y . Zhang, Z. Zhu, and D. Du, “Occformer: Dual-path transformer for vision-based 3d semantic occupancy prediction,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 9433–9443

    2023

  7. [7]

    Gaussianformer: Scene as gaussians for vision-based 3d semantic occupancy prediction,

    Y . Huang, W. Zheng, Y . Zhang, J. Zhou, and J. Lu, “Gaussianformer: Scene as gaussians for vision-based 3d semantic occupancy prediction,” arXiv preprint arXiv:2405.17429 , 2024

    arXiv 2024

  8. [8]

    M 2bev: Multi-camera joint 3d detection and seg- mentation with unified birds-eye view representation,

    E. Xie, Z. Yu, D. Zhou, J. Philion, A. Anandkumar, S. Fidler, P. Luo, and J. M. Alvarez, “M 2bev: Multi-camera joint 3d detection and seg- mentation with unified birds-eye view representation,” arXiv preprint arXiv:2204.05088, 2022

    arXiv 2022

Show all 60 references
  1. [9]

    Fast-bev: A fast and strong bird’s-eye view perception baseline,

    Y . Li, B. Huang, Z. Chen, Y . Cui, F. Liang, M. Shen, F. Liu, E. Xie, L. Sheng, W. Ouyang et al., “Fast-bev: A fast and strong bird’s-eye view perception baseline,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024

    2024

  2. [10]

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

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

    2021 arXiv

  3. [11]

    Bevdet4d: Exploit temporal cues in multi- camera 3d object detection,

    J. Huang and G. Huang, “Bevdet4d: Exploit temporal cues in multi- camera 3d object detection,” arXiv preprint arXiv:2203.17054 , 2022

    2022 arXiv

  4. [12]

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

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

    2023

  5. [13]

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

    Y . Li, H. Bao, Z. Ge, J. Yang, J. Sun, and Z. Li, “Bevstereo: Enhancing depth estimation in multi-view 3d object detection with temporal stereo,” in Proceedings of the AAAI Conference on Artificial Intelligence , vol. 37, no. 2, 2023, pp. 1486–1494

    2023

  6. [14]

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

    Y . Wang, V . C. Guizilini, T. Zhang, Y . Wang, H. Zhao, and J. Solomon, “Detr3d: 3d object detection from multi-view images via 3d-to-2d queries,” in Conference on Robot Learning . PMLR, 2022, pp. 180–191

    2022

  7. [15]

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

    Z. Li, W. Wang, H. Li, E. Xie, C. Sima, T. Lu, Y . Qiao, and J. Dai, “Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers,” in European conference on computer vision. Springer, 2022, pp. 1–18. MANUSCRIPT 12 TABLE V 3D SEM...

    2022

  8. [16]

    Bevformer v2: Adapting modern image backbones to bird’s-eye-view recognition via perspective supervision,

    C. Yang, Y . Chen, H. Tian, C. Tao, X. Zhu, Z. Zhang, G. Huang, H. Li, Y . Qiao, L. Lu et al. , “Bevformer v2: Adapting modern image backbones to bird’s-eye-view recognition via perspective supervision,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern ...

    2023

  9. [17]

    Petr: Position embedding trans- formation for multi-view 3d object detection,

    Y . Liu, T. Wang, X. Zhang, and J. Sun, “Petr: Position embedding trans- formation for multi-view 3d object detection,” in European Conference on Computer Vision . Springer, 2022, pp. 531–548

    2022

  10. [18]

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

    Y . Liu, J. Yan, F. Jia, S. Li, A. Gao, T. Wang, and X. Zhang, “Petrv2: A unified framework for 3d perception from multi-camera images,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 3262–3272

    2023

  11. [19]

    Deformable detr: De- formable transformers for end-to-end object detection,

    X. Zhu, W. Su, L. Lu, B. Li, X. Wang, and J. Dai, “Deformable detr: De- formable transformers for end-to-end object detection,” in International Conference on Learning Representations , 2021

    2021

  12. [20]

    Graph-detr3d: rethinking overlapping regions for multi-view 3d object detection,

    Z. Chen, Z. Li, S. Zhang, L. Fang, Q. Jiang, and F. Zhao, “Graph-detr3d: rethinking overlapping regions for multi-view 3d object detection,” in Proceedings of the 30th ACM International Conference on Multimedia , 2022, pp. 5999–6008

    2022

  13. [21]

    Z. Chen, Z. Chen, Z. Li, S. Zhang, L. Fang, Q. Jiang, F. Wu, and F. Zhao, “Graph-detr4d: Spatio-temporal graph modeling for multi-view 3d object TABLE IX EFFECTIVENESS OF INCREASING THE NUMBER OF 3D REFERENCE POINTS WITHIN EACH QUERY ON NU SCENES VAL SET . Model Config Nref mA...

    2024

  14. [22]

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

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

    2022 arXiv

  15. [23]

    Bev-san: Accurate bev 3d object detection via slice attention networks,

    X. Chi, J. Liu, M. Lu, R. Zhang, Z. Wang, Y . Guo, and S. Zhang, “Bev-san: Accurate bev 3d object detection via slice attention networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 17 461–17 470

    2023

  16. [24]

    Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,

    J. Philion and S. Fidler, “Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XIV 16 . Springer, 2020, pp. 194–210

    2020

  17. [25]

    Categorical depth distribution network for monocular 3d object detection,

    C. Reading, A. Harakeh, J. Chae, and S. L. Waslander, “Categorical depth distribution network for monocular 3d object detection,” in Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 8555–8564

    2021

  18. [26]

    Time will tell: New outlooks and a baseline for temporal multi-view 3d object detection,

    J. Park, C. Xu, S. Yang, K. Keutzer, K. Kitani, M. Tomizuka, and W. Zhan, “Time will tell: New outlooks and a baseline for temporal multi-view 3d object detection,” arXiv preprint arXiv:2210.02443 , 2022

    2022 arXiv

  19. [27]

    Fb- bev: Bev representation from forward-backward view transformations,

    Z. Li, Z. Yu, W. Wang, A. Anandkumar, T. Lu, and J. M. Alvarez, “Fb- bev: Bev representation from forward-backward view transformations,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 6919–6928

    2023

  20. [28]

    Bevnext: Reviving dense bev frameworks for 3d object detection,

    Z. Li, S. Lan, J. M. Alvarez, and Z. Wu, “Bevnext: Reviving dense bev frameworks for 3d object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 20 113–20 123

    2024

  21. [29]

    Exploring recurrent long-term temporal fusion for multi-view 3d perception,

    C. Han, J. Yang, J. Sun, Z. Ge, R. Dong, H. Zhou, W. Mao, Y . Peng, and X. Zhang, “Exploring recurrent long-term temporal fusion for multi-view 3d perception,” IEEE Robotics and Automation Letters , 2024

    2024

  22. [30]

    Sparse4d v2: Recurrent temporal fusion with sparse model,

    X. Lin, T. Lin, Z. Pei, L. Huang, and Z. Su, “Sparse4d v2: Recurrent temporal fusion with sparse model,” arXiv preprint arXiv:2305.14018 , 2023. MANUSCRIPT 13

    2023 arXiv

  23. [31]

    Attention is all you need,

    A. Vaswani, “Attention is all you need,” Advances in Neural Information Processing Systems, 2017

    2017

  24. [32]

    Language models are few-shot learners,

    T. B. Brown, “Language models are few-shot learners,” arXiv preprint arXiv:2005.14165, 2020

    2005 arXiv

  25. [33]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 10 012–10 022

    2021

  26. [34]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, and S. Gelly, “An image is worth 16x16 words: Transformers for image recognition at scale,” in International Conference on Learning Representations , 2021

    2021

  27. [35]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in Eu- ropean conference on computer vision . Springer, 2020, pp. 213–229

    2020

  28. [36]

    Detrs beat yolos on real-time object detection,

    Y . Zhao, W. Lv, S. Xu, J. Wei, G. Wang, Q. Dang, Y . Liu, and J. Chen, “Detrs beat yolos on real-time object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 16 965–16 974

    2024

  29. [37]

    Heightformer: Explicit height modeling without extra data for camera-only 3d object detection in bird’s eye view,

    Y . Wu, R. Li, Z. Qin, X. Zhao, and X. Li, “Heightformer: Explicit height modeling without extra data for camera-only 3d object detection in bird’s eye view,” IEEE Transactions on Image Processing , 2024

    2024

  30. [38]

    Bevheight: A robust framework for vision-based roadside 3d object detection,

    L. Yang, K. Yu, T. Tang, J. Li, K. Yuan, L. Wang, X. Zhang, and P. Chen, “Bevheight: A robust framework for vision-based roadside 3d object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 21 611–21 620

    2023

  31. [39]

    Bevheight++: Toward robust visual centric 3d object detection,

    L. Yang, T. Tang, J. Li, P. Chen, K. Yuan, L. Wang, Y . Huang, X. Zhang, and K. Yu, “Bevheight++: Toward robust visual centric 3d object detection,” arXiv preprint arXiv:2309.16179 , 2023

    2023 arXiv

  32. [40]

    Ocbev: Object-centric bev transformer for multi-view 3d object detection,

    Z. Qi, J. Wang, X. Wu, and H. Zhao, “Ocbev: Object-centric bev transformer for multi-view 3d object detection,” in 2024 International Conference on 3D Vision (3DV) . IEEE, 2024, pp. 1188–1197

    2024

  33. [41]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  34. [42]

    Feature pyramid networks for object detection,

    T.-Y . Lin, P. Doll´ar, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2017, pp. 2117–2125

    2017

  35. [43]

    Focal loss for dense object detection,

    T.-Y . Ross and G. Doll ´ar, “Focal loss for dense object detection,” in proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2980–2988

    2017

  36. [44]

    nuscenes: A multimodal dataset for autonomous driving,

    H. Caesar, V . Bankiti, A. H. Lang, S. V ora, V . E. Liong, Q. Xu, A. Krish- nan, Y . Pan, G. Baldan, and O. Beijbom, “nuscenes: A multimodal dataset for autonomous driving,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2020, pp. 11 621–11 631

    2020

  37. [45]

    Deformable convolutional networks,

    J. Dai, H. Qi, Y . Xiong, Y . Li, G. Zhang, H. Hu, and Y . Wei, “Deformable convolutional networks,” in Proceedings of the IEEE international con- ference on computer vision , 2017, pp. 764–773

    2017

  38. [46]

    Fcos3d: Fully convolutional one- stage monocular 3d object detection,

    T. Wang, X. Zhu, J. Pang, and D. Lin, “Fcos3d: Fully convolutional one- stage monocular 3d object detection,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 913–922

    2021

  39. [47]

    Decoupled weight decay regularization,

    I. Loshchilov, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101, 2017

    2017 arXiv

  40. [48]

    Sgdr: Stochastic gradient descent with warm restarts,

    I. Loshchilov and F. Hutter, “Sgdr: Stochastic gradient descent with warm restarts,” arXiv preprint arXiv:1608.03983 , 2016

    2016 arXiv

  41. [49]

    Enhancing 3d object detection with 2d detection-guided query anchors,

    H. Ji, P. Liang, and E. Cheng, “Enhancing 3d object detection with 2d detection-guided query anchors,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 21 178–21 187

    2024

  42. [50]

    Widthformer: To- ward efficient transformer-based bev view transformation,

    C. Yang, T. Lin, L. Huang, and E. J. Crowley, “Widthformer: To- ward efficient transformer-based bev view transformation,” arXiv preprint arXiv:2401.03836, 2024

    2024 arXiv

  43. [51]

    An energy and gpu- computation efficient backbone network for real-time object detection,

    Y . Lee, J.-w. Hwang, S. Lee, Y . Bae, and J. Park, “An energy and gpu- computation efficient backbone network for real-time object detection,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, 2019, pp. 0–0

    2019

  44. [52]

    Is pseudo- lidar needed for monocular 3d object detection?

    D. Park, R. Ambrus, V . Guizilini, J. Li, and A. Gaidon, “Is pseudo- lidar needed for monocular 3d object detection?” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 3142–3152

    2021

  45. [53]

    Viewpoint equivariance for multi-view 3d object detection,

    D. Chen, J. Li, V . Guizilini, R. A. Ambrus, and A. Gaidon, “Viewpoint equivariance for multi-view 3d object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 9213–9222

    2023

  46. [54]

    Lyft level 5 av dataset 2019,

    R. Kesten, M. Usman, J. Houston, T. Pandya, K. Nadhamuni, A. Ferreira, M. Yuan, B. Low, A. Jain, P. Ondruska et al. , “Lyft level 5 av dataset 2019,” 2019

    2019

  47. [55]

    MMDetection3D: OpenMMLab next-generation plat- form for general 3D object detection,

    M. Contributors, “MMDetection3D: OpenMMLab next-generation plat- form for general 3D object detection,” https://github.com/open-mmlab/ mmdetection3d, 2020

    2020

  48. [56]

    Atlas: End-to-end 3d scene reconstruction from posed images,

    Z. Murez, T. Van As, J. Bartolozzi, A. Sinha, V . Badrinarayanan, and A. Rabinovich, “Atlas: End-to-end 3d scene reconstruction from posed images,” in European conference on computer vision . Springer, 2020, pp. 414–431

    2020

  49. [57]

    Surroundocc: Multi-camera 3d occupancy prediction for autonomous driving,

    Y . Wei, L. Zhao, W. Zheng, Z. Zhu, J. Zhou, and J. Lu, “Surroundocc: Multi-camera 3d occupancy prediction for autonomous driving,” in Pro- ceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 21 729–21 740

    2023

  50. [58]

    Second: Sparsely embedded convolutional detection,

    Y . Yan, Y . Mao, and B. Li, “Second: Sparsely embedded convolutional detection,” Sensors, vol. 18, no. 10, p. 3337, 2018

    2018

  51. [59]

    Pointpillars: Fast encoders for object detection from point clouds,

    A. H. Lang, S. V ora, H. Caesar, L. Zhou, J. Yang, and O. Beijbom, “Pointpillars: Fast encoders for object detection from point clouds,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 12 697–12 705

    2019

  52. [60]

    Ssn: Shape signature networks for multi-class object detection from point clouds,

    X. Zhu, Y . Ma, T. Wang, Y . Xu, J. Shi, and D. Lin, “Ssn: Shape signature networks for multi-class object detection from point clouds,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXV 16 . Springer, 2020, pp. 581–597

    2020

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.