Fast Occupancy Network

Dong Li; Emad Barsoum; Ji Liu; Jinzhang Peng; Lu Tian; Mingjie Lu; Xingliang Huang; Yuanxian Huang

REVIEW 1 major objections 6 minor 1 cited by

Fast Occupancy Network

T0 review · 1 major / 6 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 3D occupancy prediction from cameras can be made roughly three times faster without sacrificing accuracy by lifting BEV features to voxels with a 2D deformable convolution instead of 3D attention.

desk verdict OpenOcc results hang together, but the SemanticKITTI table's own arithmetic reverses the paper's headline accuracy claim. read the letter →

arxiv 2412.07163 v1 pith:UJIYHE6S submitted 2024-12-10 cs.CV cs.AI

Mingjie Lu , Yuanxian Huang , Ji Liu , Xingliang Huang , Dong Li , Jinzhang Peng , Lu Tian , Emad Barsoum This is my paper · ORCID

classification cs.CVcs.AI

keywords occupancypredictionBEVliftingdeformableconvolutionvoxelfeaturepyramidsemanticscenecompletionautonomousdrivingperspectiveviewsupervisioninferencespeed

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 tries to establish that expensive 3D attention is not needed to turn camera-based BEV features into 3D occupancy predictions. It claims that a 2D deformable convolution applied on the BEV plane and then reshaped into height layers lifts BEV features to voxel features as accurately as 3D deformable attention, at a fraction of the cost. Two further additions, a Partial Voxel FPN that fuses multi-scale features cheaply on the horizontal plane and a train-only perspective-view segmentation loss, raise accuracy with nearly zero inference cost. On the OpenOcc benchmark the complete model surpasses OccNet by 1.7% mIoU (21.12 vs. 19.48) with a ResNet50 backbone while running at about one third of OccNet's latency. A sympathetic reader would care because this suggests camera-only occupancy prediction can become fast enough for real-time driving systems without giving up accuracy.

What carries the argument

The load-bearing mechanism is the BEV lifter: a deformable 2D convolutional layer that expands the BEV feature channels and then uses a reshape to split them into the height dimension, converting a $H \times W \times C$ BEV map into a $H \times W \times Z \times C'$ voxel feature volume. Its dynamic offsets give an adaptive receptive field, which the paper says matters because objects at different heights have different geometries. The second mechanism is the Partial Voxel FPN, which downsamples only the planar (xy) dimensions at each scale, keeps half the original features untouched, upsamples and adds the downsampled half, and applies one small 3D convolution at the coarsest scale ($50 \times 50 \times 4$) to restore height interactions. The third is the perspective-view supervision branch: a U-Net-like decoder attached to the image backbone, supervised by sparse labels made by projecting LiDAR points into each camera view, trained with focal loss and removed at inference.

What would settle it

An experiment that would settle the claim: keep the same occupancy decoder but replace the attention-based BEV construction with a depth-splatting (LSS-style) BEV builder on the OpenOcc benchmark, then compare the deformable-conv lifter against an MLP lifter. If the deformable-conv advantage shrinks to near zero, the paper's premise that BEV features contain recoverable vertical information is specific to attention-based BEV construction.

Watch

Extended reading notes

Core claim

The central claim is that BEV features already carry enough vertical information that a deformable 2D convolution over the BEV plane, followed by a channel-to-height reshape, produces voxel features comparable to 3D deformable cross-attention. The paper argues that the difference between 2D deformable attention and 3D deformable attention is mainly the number of queries and whether the height dimension is collapsed at the end, while the reference features are the same. It then packages this lifter with a Partial Voxel FPN that handles the height dimension cheaply, downsampling only the xy plane on half of the features and applying a small 3D convolution at the coarsest scale, and with a perspective-view auxiliary segmentation branch used only in training. The combined result, as reported, is a model that beats the previous state of the art on OpenOcc in both mIoU and latency, and also improves the LiDAR-segmentation-style evaluation derived from the occupancy prediction.

Load-bearing premise

The load-bearing premise is that BEV features already encode enough vertical information that a 2D deformable convolution on the BEV plane, reshaped into height layers, can recover 3D voxel features as accurately as 3D deformable attention; if this transfer fails for other BEV builders or resolutions, the reported speed-accuracy trade-off collapses.

Editorial extensions

If this is right

Any BEV perception model can be converted into an occupancy model by attaching this lifter and head; the paper demonstrates the idea on a BEVFormer-style builder, with about 22% extra latency over the plain BEVNet on ResNet50.
The cost of building voxel features no longer scales with the number of 3D queries, so larger occupancy ranges or finer voxel grids become more affordable.
The perspective-view supervision adds roughly 1.1% mIoU for free at inference, suggesting image-space supervision can compensate for the long gradient path through BEV transformation.
The Partial Voxel FPN reaches mIoU comparable to a full 3D FPN while using about one quarter of the latency, indicating most height interactions can be handled with mostly-2D operations.

Reading between the lines

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

The paper's logic implies that the 3D query stacks in other occupancy models perform redundant work if BEV features already encode height; a testable consequence is that replacing those stacks with this lifter on other BEV backbones would preserve most accuracy.
The lifter's success may depend on how the BEV is built: attention-based view transformers may imprint vertical structure, while depth-splatting (LSS-style) builders may not, so the trade-off should be re-measured on a depth-splatting BEV feature.
The perspective-view supervision idea could transfer to other BEV tasks with long gradient paths, such as online HD map construction, where a cheap image-space auxiliary loss might give a similar free gain.
Latency ratios are measured on specific accelerator hardware; on other GPUs the relative speedup could differ because deformable convolution and the small 3D convolution have different kernel optimization profiles.

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.

Desk Editor's Note

read the letter

First thing: the OpenOcc numbers are coherent, and the deformable-2D-conv lifter with partial voxel FPN is a genuinely simple efficiency idea. But the SemanticKITTI result does not survive contact with its own table. The reported mIoU for 'Ours' is 12.44; the mean of the 19 per-class IoUs in the same row is 12.27. The paper's baselines use the same 19-class mean, so under the paper's own protocol Ours is actually below OccFormer's 12.32. That reverses the 0.16% edge claimed in Section 4.4 and undercuts the abstract's 'consistently outperforms.' This isn't a rounding issue; it's a one-point swing in the opposite direction. The authors need to fix or explain this before the accuracy claims can be taken at face value.

There is also a likely copy-paste error in Table 1: the BEVNet† Res101-DCN row's listed class IoUs average to 23.78, not the reported 24.62. For the other rows the average matches, so this looks like a transcription slip rather than systematic gaming. Still, it makes the tables feel under-audited. And the intro's 'we prove that this lifting method...' is not backed by a proof anywhere; a more careful wording like 'we show empirically' would be appropriate. No code or error bars are provided, which limits how much confidence one can place in the efficiency numbers.

What the paper does well: the design is simple, the ablations in Tables 5-7 are informative, and the OpenOcc main comparison is internally consistent. If the reported latencies hold, a ~3x speedup at better mIoU on OpenOcc is a useful contribution. The paper also clearly identifies the bottleneck and explains why 2D deformable convolution is a sensible lifting mechanism. The PV supervision with projected sparse labels is a nice free-lunch trick.

The problems are concentrated in the reporting, not the architecture. A careful revision that corrects the SemanticKITTI arithmetic, fixes the Table 1 row, and tones down the 'prove' claim could make this a solid conference paper. But as is, the central 'consistent outperformance' claim is unsupported.

I'd send it to peer review because the core idea and OpenOcc evidence deserve scrutiny. But I'd flag the SemanticKITTI inconsistency as a major revision item, not a minor typo.

Referee Report

1 major / 6 minor

Summary. The paper proposes Fast Occupancy Network, a camera-based 3D semantic occupancy model whose main components are a BEV lifter based on 2D deformable convolution, a Partial Voxel FPN, and a perspective-view auxiliary segmentation loss used only at training time. On the OpenOcc/nuScenes benchmark it reports 21.12 mIoU with ResNet50 at 1.22x relative latency versus BEVNet, compared with OccNet at 19.48 mIoU and 3.43x latency, and it reports 27.22 mIoU with ResNet101-DCN. On SemanticKITTI it reports 12.44 mIoU and claims a 0.16% advantage over OccFormer. The paper argues that BEV features already contain enough vertical information for a 2D deformable convolution, applied on the BEV plane and reshaped, to replace costly 3D deformable attention.

Significance. If the OpenOcc results are reproducible, this is a useful empirical contribution: it shows that a deformable 2D convolution can lift BEV features to 3D voxel features with a large latency reduction and improved accuracy on that benchmark. The ablations in Tables 5-7 are a genuine strength because they isolate the contributions of the lifting module, PV supervision, visible mask, and Partial Voxel FPN, and they report latency proportions. The main caveat is that the SemanticKITTI evidence contains an arithmetic inconsistency in the reported mIoU that invalidates the paper's cross-dataset 'consistently outperforms' claim as currently stated. No code or machine-checked proofs are provided, so reproducibility rests on the internal consistency of the tables and the reported experimental protocol.

major comments (1)

[Table 3; Section 4.4; Table 4] The reported SemanticKITTI mIoU of 12.44 for the proposed method is not the mean of the 19 listed class IoUs. Under the same convention used for every other row in Table 3 (including OccFormer, whose listed values average to 12.32), the 'Ours' row values sum to 233.1 and average to 12.27, which is below OccFormer. Since Section 4.4 explicitly claims a 0.16% improvement over OccFormer and Table 4 repeats the 12.44 value, the central 'consistently outperforms existing methods' claim currently rests on an arithmetic error. Please correct either the per-class values or the reported mIoU, and revise the text and Table 4 accordingly.

minor comments (6)

[Table 1] The BEVNet† Res101-DCN row reports mIoU 24.62, but the 16 listed class IoUs average to approximately 23.78. This is a baseline row and does not weaken the paper's positive OpenOcc comparison, but the main comparison table should be internally consistent and the row should be corrected or footnoted.
[Section 1, last paragraph] The statement 'We prove that this lifting method is able to recover voxel information from the BEV feature' is not supported by any proof or formal argument anywhere in the manuscript; please replace 'prove' with an empirical claim or provide an actual proof.
[Section 3.2] The claim that 2D and 3D deformable cross-attention differ only in the number of queries and in whether height is collapsed at the end is imprecise, because 3D deformable attention also uses 3D reference points; rephrase to describe the actual difference accurately.
[Abstract; Section 3.2] The perspective-view branch is 'cost-free' only at inference time; it adds training-time computation and parameters, so the claim should be qualified as 'inference-cost-free.'
[Equation (3) and surrounding text] The sentence 'Here, WP Vand Wvox are represented as the weights of the weights of the PV loss and the voxel loss' contains duplicated wording and should be rewritten.
[Section 4.1] The SemanticKITTI description says there are 21 semantic classes (19 semantics, 1 free, 1 unknown), but Table 3 reports only 19 per-class columns; please clarify which classes are included in the reported mIoU.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the method is benchmarked against external OccNet/TPVFormer/OccFormer baselines and no equation reduces to a fitted input.

full rationale

The paper's central derivation is architectural and empirical: a deformable 2D convolution lifts BEV features to 3D voxel features, and the resulting occupancy predictions are evaluated on held-out OpenOcc and SemanticKITTI benchmarks. The training losses (Eq. 2 for perspective-view supervision and Eq. 3 combining PV loss with voxel loss) are supervised by externally provided LiDAR-derived labels, not by the paper's own reported mIoU values. No parameter is fitted to the comparison metrics, and the claimed improvements over OccNet, TPVFormer, and OccFormer are measured against published baselines or the paper's own reimplementations, which is standard practice. The premise that BEV features contain substantial 3D information is attributed to OccNet's BEVNet experiments, an external published result, and the paper also ablatively re-verifies it in Table 5; this is not a self-referential argument. The TPVFormer baseline shares an author with the present paper, but it is used only as a comparison target and as background, not as load-bearing justification for the method's validity, so it does not constitute circularity. The introduction's claim that 'we prove that this lifting method is able to recover voxel information' is unsupported by any formal proof and is better read as an overclaim, but it is not circular. Separately, Table 3 contains a consistency issue: the reported SemanticKITTI mIoU of 12.44 for 'Ours' does not match the mean of the listed 19 non-Road category IoUs (about 12.27), which is a correctness concern that could affect the claimed 0.16% advantage over OccFormer, but it is not a circularity because the reported metric is not defined in terms of the method's own outputs or fitted parameters. Overall, none of the seven circularity patterns is present, and the empirical core of the paper is self-contained against external benchmarks.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no invented physical entities. Its main assumptions are domain modeling choices: BEV vertical informativeness, reliability of LiDAR-projected PV labels, and sufficiency of partial multi-scale fusion. The explicit hyperparameters that are left unspecified are the loss weights in Eq. 3 and focal loss parameters.

free parameters (3)

PV loss weight W_PV = not reported
Equation 3 defines the total loss as W_PV*L_PV + W_vox*L_vox, but the paper never states W_PV or W_vox, so the training objective is not fully specified.
Voxel loss weight W_vox = not reported
Same as above; the balance between PV supervision and voxel segmentation is set by hand and not reported.
Focal loss hyperparameters (alpha, gamma) = not reported
Focal loss is used for both voxel and PV supervision, but alpha and gamma are not specified; standard defaults are likely used but unstated.

assumptions (4)

domain assumption BEV features contain enough vertical information that a 2D lifting operation can recover 3D voxel features.
Invoked in Section 3.2 as the motivation for replacing 3D deformable attention with a 2D deformable conv; the paper provides empirical support in Table 5 but no formal guarantee.
domain assumption Sparse LiDAR projections onto camera images yield reliable multi-level segmentation supervision.
Used in Section 3.2 Perspective View Supervision; assumes projected labels are correct and dense enough to train the U-Net.
domain assumption Sampling half of the voxel features and fusing only partial height layers preserves the multi-scale information that a full 3D FPN would provide.
Core design premise of Partial Voxel FPN in Section 3.2; supported only by the Table 6 ablation on one dataset.
domain assumption The OpenOcc and SemanticKITTI ground-truth annotations are correct and comparable across methods.
Dataset baseline assumption used in all experiments (Section 4.1); the paper does not audit label quality.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Fast Occupancy Network." pith.science (2026). https://pith.science/paper/UJIYHE6S

@misc{pith2026241207163,
  author       = {Pith},
  title        = {Pith review of: Fast Occupancy Network},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UJIYHE6S}},
  note         = {Machine review of arXiv:2412.07163}
}

read the original abstract

Occupancy Network has recently attracted much attention in autonomous driving. Instead of monocular 3D detection and recent bird's eye view(BEV) models predicting 3D bounding box of obstacles, Occupancy Network predicts the category of voxel in specified 3D space around the ego vehicle via transforming 3D detection task into 3D voxel segmentation task, which has much superiority in tackling category outlier obstacles and providing fine-grained 3D representation. However, existing methods usually require huge computation resources than previous methods, which hinder the Occupancy Network solution applying in intelligent driving systems. To address this problem, we make an analysis of the bottleneck of Occupancy Network inference cost, and present a simple and fast Occupancy Network model, which adopts a deformable 2D convolutional layer to lift BEV feature to 3D voxel feature and presents an efficient voxel feature pyramid network (FPN) module to improve performance with few computational cost. Further, we present a cost-free 2D segmentation branch in perspective view after feature extractors for Occupancy Network during inference phase to improve accuracy. Experimental results demonstrate that our method consistently outperforms existing methods in both accuracy and inference speed, which surpasses recent state-of-the-art (SOTA) OCCNet by 1.7% with ResNet50 backbone with about 3X inference speedup. Furthermore, our method can be easily applied to existing BEV models to transform them into Occupancy Network models.

Figures

Figures reproduced from arXiv: 2412.07163 by the authors.

**Figure 1.** Overview of our Fast Occupancy Network pipeline. design a Partial Voxel FPN structure. Unlike Voxel-FPN [13], which fuses multiscale voxel features at all levels, our approach merges multi-scale features exclusively in partial height layers. This achieves similar multi-scale fusion goals as FPN while reducing computational expenses. 3 Method 3.1 Overview The overall pipeline of our Fast Occupancy Network is shown … view at source ↗

**Figure 2.** The structure of the BEV lifter module. perspective space are weighted to obtain the corresponding features in the BEV space through a spatial cross-attention mechanism. The obtained BEV features can be used by subsequent occupancy decoders to obtain 3D voxel feature which is used to further predict semantic categories in 3D space. 3.2 Module Details BEV lifting methods. The 2D deformable cross-attention method [48]… view at source ↗

**Figure 3.** Overview of our Partial Voxel FPN. convolution as our lifting method. On the one hand, it can expand the receptive field, and on the other hand, since the geometries of objects at different heights are inconsistent, dynamic offset is necessary. In Sec 4.6, our experiments demonstrate that deformable convolution can achieve higher performance with high efficiency. Partial Voxel FPN. Inspired by FB-OCC [20], the mult… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Perspective view auxiliary FPN and loss. This structure is only utilized when training and does not influence the inference speed. Perspective View Supervision. Due to the presence of the view transformation module such as LSS [32] or BEVFormer [19], the perspective i…

**Figure 5.** Figure 5: Perspective view supervision visualization. information cannot be obtained. Therefore, we only consider the visible voxel when training to make the supervision signals more reliable. Loss Functions. Since there is a large number of empty voxels in the 3D space, we util…

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

SHTOcc: Effective 3D Occupancy Prediction with Sparse Head and Tail Voxels
cs.CV 2025-05 reject novelty 5.0 of 10

SHTOcc combines attention-based sparse voxel selection with decoupled classifier retraining for 3D occupancy prediction, reporting efficiency gains and small, partly inconsistent accuracy improvements.

Reference graph

Works this paper leans on

47 extracted references · 36 canonical work pages · cited by 1 Pith paper

[1]

In: ICCV

Behley, J., Garbade, M., Milioto, A., Quenzel, J., Behnke, S., Stachniss, C., Gall, J.: Semantickitti: A dataset for semantic scene understanding of lidar sequences. In: ICCV. pp. 9297–9307 (2019) 9, 12

work page 2019
[2]

In: CVPR

Caesar, H., Bankiti, V., Lang, A.H., Vora, S., Liong, V.E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., Beijbom, O.: nuscenes: A multimodal dataset for autonomous driving. In: CVPR. pp. 11621–11631 (2020) 9, 11

work page 2020
[3]

In: CVPR

Cao, A.Q., de Charette, R.: Monoscene: Monocular 3d semantic scene completion. In: CVPR. pp. 3991–4001 (2022) 4, 12

work page 2022
[4]

In: ECCV

Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., Zagoruyko, S.: End- to-end object detection with transformers. In: ECCV. pp. 213–229. Springer (2020) 3 Fast Occupancy Network 15

work page 2020
[5]

arXiv preprint arXiv:1706.05587 (2017) 1

Chen, L.C., Papandreou, G., Schroff, F., Adam, H.: Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587 (2017) 1

arXiv 2017
[6]

In: CVPR

Chen, X., Lin, K.Y., Qian, C., Zeng, G., Li, H.: 3d sketch-aware semantic scene completion via semi-supervised structure prior. In: CVPR. pp. 4193–4202 (2020) 2, 12

work page 2020
[7]

In: CoRL

Cheng, R., Agia, C., Ren, Y., Li, X., Bingbing, L.: S3cnet: A sparse semantic scene completion network for lidar point clouds. In: CoRL. pp. 2148–2161. PMLR (2021) 2

work page 2021
[8]

In: ICCV

Ding, W., Qiao, L., Qiu, X., Zhang, C.: Pivotnet: Vectorized pivot learning for end-to-end hd map construction. In: ICCV. pp. 3672–3682 (2023) 3

work page 2023

Show all 47 references

[9]

arXiv preprint arXiv:2203.17054 (2022) 6, 11

Huang, J., Huang, G.: Bevdet4d: Exploit temporal cues in multi-camera 3d object detection. arXiv preprint arXiv:2203.17054 (2022) 6, 11

2022 arXiv
[11]

arXiv preprint arXiv:2112.11790 (2021) 2

Huang, J., Huang, G., Zhu, Z., Ye, Y., Du, D.: Bevdet: High-performance multi- camera 3d object detection in bird-eye-view. arXiv preprint arXiv:2112.11790 (2021) 2

2021 arXiv
[12]

In: CVPR

Huang, Y., Zheng, W., Zhang, Y., Zhou, J., Lu, J.: Tri-perspective view for vision- based 3d semantic occupancy prediction. In: CVPR. pp. 9223–9232 (2023) 2, 4, 10, 11, 12

2023
[13]

Sensors20(3), 704 (2020) 4, 5

Kuang, H., Wang, B., An, J., Zhang, M., Zhang, Z.: Voxel-fpn: Multi-scale voxel feature aggregation for 3d object detection from lidar point clouds. Sensors20(3), 704 (2020) 4, 5

2020
[14]

In: CVPR

Lang, A.H., Vora, S., Caesar, H., Zhou, L., Yang, J., Beijbom, O.: Pointpillars: Fast encoders for object detection from point clouds. In: CVPR. pp. 12697–12705 (2019) 1

2019
[15]

In: CVPR

Li, J., Han, K., Wang, P., Liu, Y., Yuan, X.: Anisotropic convolutional networks for 3d semantic scene completion. In: CVPR. pp. 3351–3359 (2020) 2, 12

2020
[16]

In: ICRA

Li, Q., Wang, Y., Wang, Y., Zhao, H.: Hdmapnet: An online hd map construction and evaluation framework. In: ICRA. pp. 4628–4634. IEEE (2022) 2

2022
[17]

In: CVPR

Li, Y., Yu, Z., et al.: Voxformer: Sparse voxel transformer for camera-based 3d semantic scene completion. In: CVPR. pp. 9087–9098 (2023) 4, 6, 12

2023
[18]

In: AAAI

Li, Y., Ge, Z., Yu, G., Yang, J., Wang, Z., Shi, Y., Sun, J., Li, Z.: Bevdepth: Acquisition of reliable depth for multi-view 3d object detection. In: AAAI. vol. 37, pp. 1477–1485 (2023) 2, 3, 10, 11

2023
[19]

In: ECCV

Li, Z., Wang, W., Li, H., Xie, E., Sima, C., Lu, T., Qiao, Y., Dai, J.: Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotem- poral transformers. In: ECCV. pp. 1–18. Springer (2022) 2, 3, 5, 8, 10

2022
[20]

arXiv preprint arXiv:2307.01492 (2023) 4, 7, 13

Li, Z., Yu, Z., Austin, D., Fang, M., Lan, S., Kautz, J., Alvarez, J.M.: Fb-occ: 3d occupancy prediction based on forward-backward view transformation. arXiv preprint arXiv:2307.01492 (2023) 4, 7, 13

2023 arXiv
[21]

In: ICLR (2023) 2, 3

Liao, B., Chen, S., Wang, X., Cheng, T., Zhang, Q., Liu, W., Huang: Maptr: Structured modeling and learning for online vectorized hd map construction. In: ICLR (2023) 2, 3

2023
[22]

arXiv preprint arXiv:2308.05736 (2023) 2

Liao, B., Chen, S., Zhang, Y., Jiang, B., Zhang, Q., Liu, W., Huang, C., Wang, X.: Maptrv2: An end-to-end framework for online vectorized hd map construction. arXiv preprint arXiv:2308.05736 (2023) 2

2023 arXiv
[23]

In: CVPR

Lin, T.Y., Dollár, P., Girshick, R., He, K., Hariharan, B., Belongie, S.: Feature pyramid networks for object detection. In: CVPR. pp. 2117–2125 (2017) 4 16 Lu et al

2017
[24]

In: ICCV

Lin, T.Y., Goyal, P., Girshick, R., He, K., Dollár, P.: Focal loss for dense object detection. In: ICCV. pp. 2980–2988 (2017) 9

2017
[25]

In: ECCV

Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.Y., Berg, A.C.: Ssd: Single shot multibox detector. In: ECCV. pp. 21–37. Springer (2016) 1

2016
[26]

In: ICML

Liu, Y., Yuan, T., Wang, Y., Wang, Y., Zhao, H.: Vectormapnet: End-to-end vec- torized hd map learning. In: ICML. pp. 22352–22369. PMLR (2023) 2

2023
[27]

In: ICCV

Liu, Y., Yan, J., Jia, F., Li, S., Gao, A., Wang, T., Zhang, X.: Petrv2: A unified framework for 3d perception from multi-camera images. In: ICCV. pp. 3262–3272 (2023) 2

2023
[28]

arXiv preprint arXiv:2302.13540 (2023) 4

Miao, R., Liu, W., Chen, M., Gong, Z., Xu, W., Hu, C., Zhou, S.: Oc- cdepth: A depth-aware method for 3d semantic scene completion. arXiv preprint arXiv:2302.13540 (2023) 4

2023 arXiv
[29]

arXiv preprint arXiv:2006.11436 (2020) 3

Ng, M.H., Radia, K., Chen, J., Wang, D., Gog, I., Gonzalez, J.E.: Bev-seg: Bird’s eye view semantic segmentation using geometry and semantic point cloud. arXiv preprint arXiv:2006.11436 (2020) 3

2020 arXiv
[30]

RA-L5(3), 4867–4873 (2020) 3

Pan, B., Sun, J., Leung, H.Y.T., Andonian, A., Zhou, B.: Cross-view semantic segmentation for sensing surroundings. RA-L5(3), 4867–4873 (2020) 3

2020
[31]

In: WACV

Peng, L., Chen, Z., Fu, Z., Liang, P., Cheng, E.: Bevsegformer: Bird’s eye view se- mantic segmentation from arbitrary camera rigs. In: WACV. pp. 5935–5943 (2023) 2

2023
[32]

In: ECCV (2020) 3, 8

Philion, J., Fidler, S.: Lift, splat, shoot: Encoding images from arbitrary camera rigs by implicitly unprojecting to 3d. In: ECCV (2020) 3, 8

2020
[33]

In: CVPR

Qu, Z., Jin, H., Zhou, Y., Yang, Z., Zhang, W.: Focus on local: Detecting lane marker from bottom up via key point. In: CVPR. pp. 14122–14130 (2021) 1

2021
[34]

arXiv preprint arXiv:1804.02767 (2018) 1

Redmon, J., Farhadi, A.: Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767 (2018) 1

2018 arXiv
[35]

Roldao, L., de Charette, R., Verroust-Blondet, A.: Lmscnet: Lightweight multiscale 3d semantic completion. In: 3DV. pp. 111–119. IEEE (2020) 2, 12

2020
[36]

In: CVPR

Song, S., Yu, F., Zeng, A., Chang, A.X., Savva, M., Funkhouser, T.: Semantic scene completion from a single depth image. In: CVPR. pp. 1746–1754 (2017) 2

2017
[37]

In: CVPR

Tabelini, L., Berriel, R., Paixao, T.M., Badue, C., De Souza, A.F., Oliveira-Santos, T.: Keep your eyes on the lane: Real-time attention-guided lane detection. In: CVPR. pp. 294–302 (2021) 1

2021
[38]

In: ICML

Tan, M., Le, Q.: Efficientnet: Rethinking model scaling for convolutional neural networks. In: ICML. pp. 6105–6114. PMLR (2019) 10

2019
[39]

In: ICCV

Tong, W., Sima, C., Wang, T., Chen, L., Wu, S., Deng, H., Gu, Y., Lu, L., Luo, P., Lin, D., et al.: Scene as occupancy. In: ICCV. pp. 8406–8415 (2023) 2, 3, 4, 6, 9, 10, 11, 12

2023
[40]

In: CVPR

Vora, S., Lang, A.H., Helou, B., Beijbom, O.: Pointpainting: Sequential fusion for 3d object detection. In: CVPR. pp. 4604–4612 (2020) 1

2020
[41]

In: ICCV

Wang, T., Zhu, X., Pang, J., Lin, D.: Fcos3d: Fully convolutional one-stage monoc- ular 3d object detection. In: ICCV. pp. 913–922 (2021) 10

2021
[42]

arXiv preprint arXiv:2306.10013 (2023) 2, 4, 6

Wang, Y., Chen, Y., Liao, X., Fan, L., Zhang, Z.: Panoocc: Unified occu- pancy representation for camera-based 3d panoptic segmentation. arXiv preprint arXiv:2306.10013 (2023) 2, 4, 6

2023 arXiv
[43]

In: ITSC

Xu, S., Zhou, D., Fang, J., Yin, J., Bin, Z., Zhang, L.: Fusionpainting: Multimodal fusion with adaptive attention for 3d object detection. In: ITSC. pp. 3047–3054. IEEE (2021) 1

2021
[44]

In: AAAI

Yan, X., Gao, J., Li, J., Zhang, R., Li, Z., Huang, R., Cui, S.: Sparse single sweep lidar point cloud segmentation via learning contextual shape priors from scene completion. In: AAAI. vol. 35, pp. 3101–3109 (2021) 12 Fast Occupancy Network 17

2021
[45]

In: CVPR

Yin, T., Zhou, X., Krahenbuhl, P.: Center-based 3d object detection and tracking. In: CVPR. pp. 11784–11793 (2021) 1

2021
[46]

In: ECCV

Yuan, Y., Chen, X., Wang, J.: Object-contextual representations for semantic seg- mentation. In: ECCV. pp. 173–190. Springer (2020) 1

2020
[47]

arXiv preprint arXiv:2304.05316 (2023) 2, 4, 11, 12

Zhang, Y., Zhu, Z., Du, D.: Occformer: Dual-path transformer for vision-based 3d semantic occupancy prediction. arXiv preprint arXiv:2304.05316 (2023) 2, 4, 11, 12

2023 arXiv
[48]

arXiv preprint arXiv:2010.04159 (2020) 6

Zhu, X., Su, W., Lu, L., Li, B., Wang, X., Dai, J.: Deformable detr: Deformable transformers for end-to-end object detection. arXiv preprint arXiv:2010.04159 (2020) 6

2020 arXiv

Pith tools

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

[1] [1]

In: ICCV

Behley, J., Garbade, M., Milioto, A., Quenzel, J., Behnke, S., Stachniss, C., Gall, J.: Semantickitti: A dataset for semantic scene understanding of lidar sequences. In: ICCV. pp. 9297–9307 (2019) 9, 12

work page 2019

[2] [2]

In: CVPR

Caesar, H., Bankiti, V., Lang, A.H., Vora, S., Liong, V.E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., Beijbom, O.: nuscenes: A multimodal dataset for autonomous driving. In: CVPR. pp. 11621–11631 (2020) 9, 11

work page 2020

[3] [3]

In: CVPR

Cao, A.Q., de Charette, R.: Monoscene: Monocular 3d semantic scene completion. In: CVPR. pp. 3991–4001 (2022) 4, 12

work page 2022

[4] [4]

In: ECCV

Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., Zagoruyko, S.: End- to-end object detection with transformers. In: ECCV. pp. 213–229. Springer (2020) 3 Fast Occupancy Network 15

work page 2020

[5] [5]

arXiv preprint arXiv:1706.05587 (2017) 1

Chen, L.C., Papandreou, G., Schroff, F., Adam, H.: Rethinking atrous convolution for semantic image segmentation. arXiv preprint arXiv:1706.05587 (2017) 1

arXiv 2017

[6] [6]

In: CVPR

Chen, X., Lin, K.Y., Qian, C., Zeng, G., Li, H.: 3d sketch-aware semantic scene completion via semi-supervised structure prior. In: CVPR. pp. 4193–4202 (2020) 2, 12

work page 2020

[7] [7]

In: CoRL

Cheng, R., Agia, C., Ren, Y., Li, X., Bingbing, L.: S3cnet: A sparse semantic scene completion network for lidar point clouds. In: CoRL. pp. 2148–2161. PMLR (2021) 2

work page 2021

[8] [8]

In: ICCV

Ding, W., Qiao, L., Qiu, X., Zhang, C.: Pivotnet: Vectorized pivot learning for end-to-end hd map construction. In: ICCV. pp. 3672–3682 (2023) 3

work page 2023