Pith. sign in

REVIEW 3 major objections 7 minor 41 references

Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detection

T0 review · 3 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 A network that completes per-object occupancy volumes from noisy LiDAR tracklets improves 3D detection on the Waymo Open Dataset, with the largest gains on distant vehicles.

desk verdict Worth a serious look: the detection gains are real and measured on a public benchmark, but the shape-completion IoU numbers rest on self-generated pseudo-labels that never receive an external validity check. read the letter →

arxiv 2412.05154 v1 pith:KQWBLFSE submitted 2024-12-06 cs.CV cs.AI

Chaoda Zheng , Feng Wang , Naiyan Wang , Shuguang Cui , Zhen Li This is my paper · ORCID
classification cs.CVcs.AI
keywords object-centricoccupancy3DobjectdetectioncompletionimplicitshapedecoderLiDARperceptionautonomousdrivingtemporalaggregationWaymoOpenDataset
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 3D bounding boxes are too coarse for autonomous driving perception, because they treat the entire interior of a box as occupied. It introduces object-centric occupancy — a per-object voxel grid that records which parts of the object are actually filled — and builds the first object-centric occupancy dataset from scratch using an automated pipeline over the Waymo Open Dataset. On the algorithmic side, it trains a sequence-based occupancy completion network whose implicit shape decoder generates a complete occupancy volume for each tracked proposal, even when the proposals are noisy. The paper's central demonstration is that the resulting occupancy features improve 3D detection: L1 AP for CenterPoint rises from 73.2 to 81.8, FSD reaches 83.3, and the largest gains appear on vehicles beyond 50 meters.

What carries the argument

The load-bearing mechanism is the object-centric occupancy volume defined in the object's own coordinate system. It is produced by an automated pipeline that aggregates LiDAR points inside ground-truth boxes across a tracked sequence, voxelizes them at 0.2 m resolution, and labels unoccupied voxels as free or unobserved by comparing each voxel center's range value with the raw range image. On the network side, a dual-branch RoI encoder represents each proposal in both local and global coordinates, a causal transformer aggregates features across history frames, and an implicit shape decoder — an MLP $D(z, q)$ that maps a latent shape code and a continuous query position to an occupancy probability — generates occupancy volumes of arbitrary, dynamic size. The implicit decoder is what allows the network to be trained end-to-end with detection, to extrapolate occupancy outside the original proposal, and to avoid the cost of dense or sparse convolution on variable-sized volumes.

What would settle it

Take a sample of Waymo vehicles with predicted occupancy volumes and compare them against dense ground truth obtained from a high-resolution scanner or manual voxel labeling. If the range-image based free/unobserved labeling disagrees substantially with ray casting or with the ground truth, the training labels are corrupted, and the reported shape IoUs and the detection gains would need to be re-measured against corrected labels.

Watch

Extended reading notes

Core claim

The central claim is that a complete object-centric occupancy volume, decoded from a latent shape embedding by an implicit function, is a more informative geometric description of a detected object than a bounding box, and that this extra geometry feeds back into detection. The paper shows that its network predicts full vehicle shapes from noisy, tracked LiDAR proposals with shape IoUs well above the voxelized-input baseline, and that fusing the learned shape latent with detection features yields consistent AP improvements across two base detectors, with and without retraining on the new detector's tracklets. The range breakdown makes the mechanism concrete: on the [50,+inf) band, L2 mAP improves by 13.94 points for CenterPoint and 7.81 for FSD, exactly where single-frame LiDAR is sparsest and box estimates are least reliable.

Load-bearing premise

The automatic annotation pipeline that labels free, occupied, and unobserved voxels must be accurate: it assumes the ground-truth boxes are reliable, that vehicles are rigid so points from different frames can be aligned, and that the range-image comparison correctly identifies which empty voxels are truly free.

Editorial extensions

If this is right

  • Downstream planning could treat each detected object as its actual occupied volume rather than a filled box, which matters for irregular shapes such as cranes or vehicles with protruding loads.
  • The shape prior transfers across detectors: a model trained on CenterPoint and FSD tracklets improves FSDv2 without retraining, suggesting object-centric occupancy is a detector-agnostic representation.
  • Longer history improves both shape IoU and detection AP, and the online full-history setting nearly matches the offline future-aware setting, so the representation is compatible with real-time constraints.
  • The largest detection gains at the 50+ m range indicate that occupancy completion supplies geometry that raw point density cannot, which could extend the effective range of LiDAR perception.

Reading between the lines

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

  • The same annotation recipe could be applied to other large-scale LiDAR datasets with box annotations, producing object-centric occupancy pretraining data without any manual shape labeling.
  • Because the implicit decoder is queried at continuous coordinates, an object-centric occupancy representation could be rendered at arbitrary resolutions at inference time, letting a downstream system trade detail for compute per scene.
  • A direct validation of the occlusion-reasoning shortcut would be to compare its free/unobserved decisions with full ray casting on a subset of frames; disagreement there would pin down the main source of label noise in the dataset.
  • The consistent detection gains suggest occupancy completion could serve as a self-supervised training signal for detectors, since the shape latent learns geometry that helps localize sparse, distant objects.
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

3 major / 7 minor

Summary. The paper introduces object-centric occupancy as a representation to supplement 3D bounding boxes for driving perception. It constructs an automated annotation pipeline on the Waymo Open Dataset that aggregates LiDAR points inside GT boxes over time, voxelizes them in the object coordinate system, and uses range-image comparison to label unoccupied voxels as free or unobserved. The proposed network takes noisy object tracklet proposals, encodes each RoI with local and global branches, applies causal temporal attention, and uses an implicit MLP shape decoder to produce occupancy volumes of arbitrary size; the latent is also fused with the global feature for detection refinement. On WOD validation, the method reports higher shape-completion IoU than direct point accumulation under GT, noisy, and detector-generated tracklets, and improves CenterPoint and FSD detection AP by large margins, especially beyond 50 m. Ablations and sequence-length studies support the architecture choices.

Significance. If the reported results hold, this is a useful contribution: it defines a practical object-centric occupancy task, provides a dataset-generation pipeline, demonstrates that an implicit decoder can be integrated with single-frame detectors to yield large, range-dependent gains on a standard benchmark, and releases code. I credit the authors for using official Waymo metrics, for testing robustness on noisy tracklets, for the FSDv2 no-retraining generalization result, and for the direct 'No Occ. Dec.' ablation. The principal risk is that all shape-completion ground truth and evaluation targets come from the same automatic pipeline, with no independent check; this affects the interpretation of Table 1 and, to a lesser degree, the occupancy-based component of the detection gains.

major comments (3)
  1. [Section 3.2, Eq. (6), Table 1] The central shape-completion claim rests on occupancy labels generated entirely by the authors' pipeline from Waymo GT boxes and range images. The occupancy loss in Eq. (5) and every IoU in Table 1 are measured against these same labels, so systematic pipeline errors—for example, in the rigid-body alignment used to aggregate points, in the assumption that GT boxes are accurate enough to collect an object's points, or in the free/unobserved decision rule of Eq. (6)—would corrupt training and evaluation coherently and could inflate the apparent completion accuracy. I ask for a validation study on a subset of objects (manual labeling, CAD alignment, or independent reconstruction) and for a sensitivity analysis of pipeline parameters (voxel size, range threshold, and bbox noise), and I recommend that the abstract and conclusion soften 'accurately predicts the complete object-centric occupancy' until such validation is provided.
  2. [Section 5.3, Fig. 5] The IoU protocol does not actually measure completion of unobserved regions. Unobserved GT voxels are ignored during IoU computation, and voxels outside the predicted RoI are assigned 'free,' so the metric primarily scores agreement with the pipeline's observed/free labels and with RoI coverage rather than with the complete 3D shape. This is why the 'unobserved' class matters: a prediction that regards a truly occupied occluded region as free incurs no penalty. Please report an additional metric that includes unobserved voxels, or a separate evaluation on voxels that are unobserved in the input sequence but occupied in a validated dense reconstruction; without this, the term 'complete occupancy' is not supported by Table 1.
  3. [Section 5.5, Table 4] The 'No Occ. Dec.' ablation removes the occupancy decoder and its loss simultaneously, so the detection drop from 82.80 to 81.10 L1 AP could be due to removing an auxiliary training signal or part of the model capacity rather than to occupancy semantics. To substantiate the statement that occupancy features 'significantly enhance' detection, I recommend a control with a comparably-capable auxiliary head trained on a non-occupancy objective (for example, point-density or interior-mask prediction) and a sweep of the occupancy-loss weight. The range breakdown in Table 3 is consistent with the authors' explanation, but it does not identify the mechanism by itself.
minor comments (7)
  1. [Table 3] The caption 'L2 mAP' is inaccurate; Table 3 reports L2 AP/APH at IoU 0.7, not mean average precision. Please correct the label and any corresponding text.
  2. [Section 4.3 / Figure 4] 'CausualAttn' is a typo for 'CausalAttn'; also please define the dimensions of gamma(T) and phi(B) and state how the bbox embedding is computed before being added to the global features.
  3. [Table 2 / Section 5.4] The '[-p,+f]' frame notation is never defined, and Table 2 mixes single-frame, multi-frame, online, and offline entries (for example, 3D-MAN and MVF++ use future frames). Please define the notation and scope the 'state-of-the-art online detection' claim to methods with the same frame-access assumptions.
  4. [Appendix A.1 / Figure 7] There are small textual typos: 'traversed trough' should be 'traversed through' in Appendix A.1, and 'colume' should be 'column' in the Figure 7 caption.
  5. [Section 2.1] The statement that SOP 'only needs to predict the occupancy for visible regions' is imprecise; common SOP benchmarks expect a full scene occupancy volume, not only visible regions. Please rephrase to distinguish visible-input conditions from output scope.
  6. [Abstract / Limitations] The evaluation is restricted to vehicles, yet the title and abstract make unqualified claims about 'objects' and 'flexible 3D perception.' I suggest explicitly stating the vehicle-only scope in the abstract or adding a sentence in the conclusion that all experiments concern rigid objects.
  7. [Section 5.1] Please clarify how the fixed-length (32-frame) training tracklets are formed from FSD and CenterPoint proposals, whether padding is applied at the beginning or end, and how 'all history frames' is interpreted for the [-infinity,0] inference setting.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the main claims are empirical, benchmark-tested, and do not reduce to their inputs by construction.

full rationale

The paper's two central claims—(1) that the proposed network accurately completes object-centric occupancy from noisy proposals, and (2) that the learned occupancy features improve 3D detection—are empirical claims evaluated on the Waymo Open Dataset. Detection results in Tables 2-4 use the official WOD AP/APH metrics against ground-truth boxes, which are external to the proposed method. Shape-completion IoU in Table 1 is measured on the Waymo validation set against occupancy labels produced by the authors' automated pipeline. This introduces a label-quality concern: the same pipeline generates both the training supervision and the evaluation ground truth, and no independent validation (e.g., manual segmentation or CAD models) is provided. However, this is a data-quality and external-validity limitation, not a circular derivation. Nothing in the paper defines a quantity in terms of the target result, and no fitted parameter is renamed as a prediction. The network is trained on one set of sequences and evaluated on held-out sequences and under noisy/detector-generated tracklets, so the reported IoU and detection gains are not guaranteed by construction. The self-citations to FSD, LiDAR R-CNN, and ImmortalTracker are used as architectural components or proposal/tracklet generators; they are load-bearing engineering choices but are not invoked as a uniqueness theorem or as the sole justification for the central claim. The acknowledged rigid-body limitation narrows the scope to vehicles but does not make the derivation circular. Overall, the paper's logic is self-contained and benchmark-anchored, with only mild dependencies on author-generated labels and prior work, neither of which rises to circularity.

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

The central empirical claim rests on the quality of the auto-generated occupancy labels, the assumption that temporal context can infer complete shapes from partial LiDAR, and a set of hand-chosen hyperparameters that are not learned from data. No new physical or theoretical entities are introduced; the latent shape embedding is a learned representation whose predictions are directly evaluated against occupancy labels.

free parameters (6)
  • Voxel size for object-centric occupancy = 0.2 m
    Chosen to balance fine-grained object detail against computational cost (Section 5.2, Appendix A.4); determines the resolution of both annotations and evaluation.
  • Training sequence length = 32 frames (padded or cut)
    Set as a compromise between performance and cost after the length study in Table 5; inference is flexible but training uses 32.
  • Position query sampling budget = 1024 queries per RoI, 512 occupied / 512 free
    Balanced sampling strategy in Section 5.1 to avoid bias in the binary cross-entropy occupancy loss.
  • Loss weights = lambda_det = 2, lambda_score = 1
    Hand-set hyperparameters in Eq. (5) that balance occupancy, box, and objectness losses; not optimized or justified beyond empirical choice.
  • Transformer configuration = 3 layers, 4 heads, hidden dim 512
    Architecture capacity choice for the causal temporal aggregation (Appendix A.3).
  • Optimization schedule = Adam, lr=1e-4, batch size 8, 24 epochs, cosine annealing
    Training recipe in Appendix A.3; affects final numbers but is standard practice.
assumptions (4)
  • domain assumption The automated annotation pipeline produces accurate ground-truth occupancy for rigid vehicles.
    Section 3.2 and Eq. (6). The entire shape completion evaluation depends on the pseudo-labels being correct; no external validation of the generated occupancy is provided.
  • domain assumption Vehicle objects are rigid and can be aligned over time by their annotated bbox poses.
    Section 3.2 and the Limitations paragraph: deformable objects are excluded because the rigid-body assumption would break the aggregated shape.
  • domain assumption GT bbox annotations and tracklet associations are accurate enough to define object-centric coordinate systems.
    The pipeline gathers points within GT boxes and relies on ImmortalTracker, FSD, or CenterPoint tracklets for input sequences (Sections 3.2 and 5.1).
  • domain assumption The learned implicit function generalizes over continuous query positions inside and slightly outside RoIs.
    Section 4.1 and the Ours-E result in Table 1; extrapolation beyond the RoI is used to claim flexibility, but there is no guarantee of correctness outside the training distribution.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detection." pith.science (2026). https://pith.science/paper/KQWBLFSE

@misc{pith2026241205154,
  author       = {Pith},
  title        = {Pith review of: Towards Flexible 3D Perception: Object-Centric Occupancy Completion Augments 3D Object Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KQWBLFSE}},
  note         = {Machine review of arXiv:2412.05154}
}
read the original abstract

While 3D object bounding box (bbox) representation has been widely used in autonomous driving perception, it lacks the ability to capture the precise details of an object's intrinsic geometry. Recently, occupancy has emerged as a promising alternative for 3D scene perception. However, constructing a high-resolution occupancy map remains infeasible for large scenes due to computational constraints. Recognizing that foreground objects only occupy a small portion of the scene, we introduce object-centric occupancy as a supplement to object bboxes. This representation not only provides intricate details for detected objects but also enables higher voxel resolution in practical applications. We advance the development of object-centric occupancy perception from both data and algorithm perspectives. On the data side, we construct the first object-centric occupancy dataset from scratch using an automated pipeline. From the algorithmic standpoint, we introduce a novel object-centric occupancy completion network equipped with an implicit shape decoder that manages dynamic-size occupancy generation. This network accurately predicts the complete object-centric occupancy volume for inaccurate object proposals by leveraging temporal information from long sequences. Our method demonstrates robust performance in completing object shapes under noisy detection and tracking conditions. Additionally, we show that our occupancy features significantly enhance the detection results of state-of-the-art 3D object detectors, especially for incomplete or distant objects in the Waymo Open Dataset.

Figures

Figures reproduced from arXiv: 2412.05154 by the authors.

Figure 1
Figure 1. Bounding Box vs. Occupancy. Oc￾cupancy can better represent the crane’s shape than the bounding box. In autonomous driving, accurate and robust 3D scene perception is crucial for safe and efficient navigation. Conventional perception systems pri￾marily adopt 3D object bounding boxes as the per￾ception representation [25, 6, 15, 16]. However, the limitations of 3D bounding boxes (bboxes) are be￾coming increasingly pr… view at source ↗
Figure 2
Figure 2. Generating occupancy from LiDAR scans is non-trivial for foreground objects due to sparsity and detection drifts. Considering the limitation of 3D bounding boxes, occupancy representation has emerged as a promis￾ing alternative for 3D scene perception [30, 29, 33]. As shown in [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Transforming these jagged object voxels to the object’s coordinate system inevitably leads [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Architecture overview. The network takes a noisy object sequence as input and outputs [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Illustration for occupancy evaluation. Evaluation Metrics. For shape completion, we adopt the widely-used intersection-over-union (IoU) to evaluate the quality of the predicted occupancy volumes. Due to the object-centric nature of our method, we cannot calculate the I…
Figure 6
Figure 6. Figure 6: Our object-centric occupancy annotation pipeline. [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Visualization of our object-centric occupancy annotations. The first colume shows the [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Single-branch model architecture. Query pos. Pred. occ. BBox Global Feats. 𝒵! 𝑡" 𝑡# 𝑡$ 𝑡% … [,] MLP Causal Attention MLP Score 0.95 MLP Repeat [,] Bbox Head 𝑆ℎ𝑎𝑝𝑒 𝐸𝑚𝑏. 𝒵& 𝑆ℎ𝑎𝑝𝑒 𝐸𝑚𝑏. 𝒵 ℇ𝒈𝒍𝒐𝒃𝒂𝒍 ℇ𝒍𝒐𝒄𝒂𝒍 𝓓 Occ Head Global Feats. 𝒵!′ ℇ𝒈𝒍𝒐𝒃𝒂𝒍 Sample [PITH_FULL_IMAGE:figures/…
Figure 9
Figure 9. Figure 9: Architecture of using explicit occupancy for detection. [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Visualization of the object-centric occupancy prediction. Different rows denote different [PITH_FULL_IMAGE:figures/full_fig_p016_10.png]
Figure 11
Figure 11. Figure 11: The renderings of predicted occupancy decoded from the shape codes for common vehicles. [PITH_FULL_IMAGE:figures/full_fig_p017_11.png]
Figure 12
Figure 12. Figure 12: The renderings of complex vehicles. Each row shows the rendering, the corresponding [PITH_FULL_IMAGE:figures/full_fig_p017_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 36 canonical work pages

  1. [1]

    Scan2cad: Learning cad model alignment in rgb-d scans

    Armen Avetisyan, Manuel Dahnert, Angela Dai, Manolis Savva, Angel X Chang, and Matthias Nießner. Scan2cad: Learning cad model alignment in rgb-d scans. In IEEE Conf. Comput. Vis. Pattern Recog., pages 2614–2623, 2019

    work page 2019

  2. [2]

    nuscenes: A multimodal dataset for autonomous driving

    Holger Caesar, Varun Bankiti, Alex H Lang, Sourabh V ora, Venice Erin Liong, Qiang Xu, Anush Krishnan, Yu Pan, Giancarlo Baldan, and Oscar Beijbom. nuscenes: A multimodal dataset for autonomous driving. In IEEE Conf. Comput. Vis. Pattern Recog., pages 11621–11631, 2020

    work page 2020

  3. [3]

    Mppnet: Multi-frame feature intertwining with proxy points for 3d temporal object detection

    Xuesong Chen, Shaoshuai Shi, Benjin Zhu, Ka Chun Cheung, Hang Xu, and Hongsheng Li. Mppnet: Multi-frame feature intertwining with proxy points for 3d temporal object detection. In Eur. Conf. Comput. Vis., pages 680–697. Springer, 2022

    work page 2022

  4. [4]

    V oxelnext: Fully sparse voxelnet for 3d object detection and tracking

    Yukang Chen, Jianhui Liu, Xiangyu Zhang, Xiaojuan Qi, and Jiaya Jia. V oxelnext: Fully sparse voxelnet for 3d object detection and tracking. In IEEE Conf. Comput. Vis. Pattern Recog., pages 21674–21683, 2023

    work page 2023

  5. [5]

    Spconv: Spatially sparse convolution library

    Spconv Contributors. Spconv: Spatially sparse convolution library. https://github.com/ traveller59/spconv, 2022

    2022

  6. [6]

    Fully sparse 3d object detection

    Lue Fan, Feng Wang, Naiyan Wang, and ZHAO-XIANG ZHANG. Fully sparse 3d object detection. Advances in Neural Information Processing Systems, 35:351–363, 2022

    work page 2022

  7. [7]

    Fsd v2: Improving fully sparse 3d object detection with virtual voxels

    Lue Fan, Feng Wang, Naiyan Wang, and Zhaoxiang Zhang. Fsd v2: Improving fully sparse 3d object detection with virtual voxels. arXiv preprint arXiv:2308.03755, 2023

    arXiv 2023

  8. [8]

    Once detected, never lost: Surpassing human performance in offline lidar based 3d object detection

    Lue Fan, Yuxue Yang, Yiming Mao, Feng Wang, Yuntao Chen, Naiyan Wang, and Zhaoxiang Zhang. Once detected, never lost: Surpassing human performance in offline lidar based 3d object detection. In Int. Conf. Comput. Vis., pages 19820–19829, 2023

    work page 2023

Show all 41 references
  1. [9]

    Super sparse 3d object detection

    Lue Fan, Yuxue Yang, Feng Wang, Naiyan Wang, and Zhaoxiang Zhang. Super sparse 3d object detection. IEEE Trans. Pattern Anal. Mach. Intell., 2023

    2023

  2. [10]

    3d semantic segmentation with submanifold sparse convolutional networks

    Benjamin Graham, Martin Engelcke, and Laurens van der Maaten. 3d semantic segmentation with submanifold sparse convolutional networks. IEEE Conf. Comput. Vis. Pattern Recog., 2018

    2018

  3. [11]

    Selfocc: Self- supervised vision-based 3d occupancy prediction

    Yuanhui Huang, Wenzhao Zheng, Borui Zhang, Jie Zhou, and Jiwen Lu. Selfocc: Self- supervised vision-based 3d occupancy prediction. In IEEE Conf. Comput. Vis. Pattern Recog., pages 19946–19956, 2024

    2024

  4. [12]

    Tri-perspective view for vision-based 3d semantic occupancy prediction

    Yuanhui Huang, Wenzhao Zheng, Yunpeng Zhang, Jie Zhou, and Jiwen Lu. Tri-perspective view for vision-based 3d semantic occupancy prediction. In IEEE Conf. Comput. Vis. Pattern Recog., pages 9223–9232, 2023

    2023

  5. [13]

    3d distance fields: A survey of techniques and applications

    Mark W Jones, J Andreas Baerentzen, and Milos Sramek. 3d distance fields: A survey of techniques and applications. IEEE Trans. Vis. Comput. Graph., 12(4):581–599, 2006

    2006

  6. [14]

    Modar: Using motion forecasting for 3d object detection in point cloud sequences

    Yingwei Li, Charles R Qi, Yin Zhou, Chenxi Liu, and Dragomir Anguelov. Modar: Using motion forecasting for 3d object detection in point cloud sequences. In IEEE Conf. Comput. Vis. Pattern Recog., pages 9329–9339, 2023

    2023

  7. [15]

    Lidar r-cnn: An efficient and universal 3d object detector

    Zhichao Li, Feng Wang, and Naiyan Wang. Lidar r-cnn: An efficient and universal 3d object detector. In IEEE Conf. Comput. Vis. Pattern Recog., pages 7546–7555, 2021. 11

    2021

  8. [16]

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

    Zhiqi Li, Wenhai Wang, Hongyang Li, Enze Xie, Chonghao Sima, Tong Lu, Yu Qiao, and Jifeng Dai. Bevformer: Learning bird’s-eye-view representation from multi-camera images via spatiotemporal transformers. In Eur. Conf. Comput. Vis.Springer, 2022

    2022

  9. [17]

    Learning a room with the occ-sdf hybrid: Signed distance function mingled with occupancy aids scene representation

    Xiaoyang Lyu, Peng Dai, Zizhang Li, Dongyu Yan, Yi Lin, Yifan Peng, and Xiaojuan Qi. Learning a room with the occ-sdf hybrid: Signed distance function mingled with occupancy aids scene representation. In Int. Conf. Comput. Vis., pages 8940–8950, 2023

    2023

  10. [18]

    Nerf: Representing scenes as neural radiance fields for view synthesis

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoor- thi, and Ren Ng. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65(1):99–106, 2021

    2021

  11. [19]

    Deepsdf: Learning continuous signed distance functions for shape representation

    Jeong Joon Park, Peter Florence, Julian Straub, Richard Newcombe, and Steven Lovegrove. Deepsdf: Learning continuous signed distance functions for shape representation. In Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 165–174, 2019

    2019

  12. [20]

    Pointnet: Deep learning on point sets for 3d classification and segmentation

    Charles R Qi, Hao Su, Kaichun Mo, and Leonidas J Guibas. Pointnet: Deep learning on point sets for 3d classification and segmentation. In IEEE Conf. Comput. Vis. Pattern Recog., pages 652–660, 2017

    2017

  13. [21]

    Offboard 3d object detection from point cloud sequences

    Charles R Qi, Yin Zhou, Mahyar Najibi, Pei Sun, Khoa V o, Boyang Deng, and Dragomir Anguelov. Offboard 3d object detection from point cloud sequences. In IEEE Conf. Comput. Vis. Pattern Recog., pages 6134–6144, 2021

    2021

  14. [22]

    Frodo: From detections to 3d objects

    Martin Runz, Kejie Li, Meng Tang, Lingni Ma, Chen Kong, Tanner Schmidt, Ian Reid, Lourdes Agapito, Julian Straub, Steven Lovegrove, et al. Frodo: From detections to 3d objects. In IEEE Conf. Comput. Vis. Pattern Recog., pages 14720–14729, 2020

    2020

  15. [23]

    Slam++: Simultaneous localisation and mapping at the level of objects

    Renato F Salas-Moreno, Richard A Newcombe, Hauke Strasdat, Paul HJ Kelly, and Andrew J Davison. Slam++: Simultaneous localisation and mapping at the level of objects. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1352–1359, 2013

    2013

  16. [24]

    Pv-rcnn++: Point-voxel feature set abstraction with local vector representa- tion for 3d object detection

    Shaoshuai Shi, Li Jiang, Jiajun Deng, Zhe Wang, Chaoxu Guo, Jianping Shi, Xiaogang Wang, and Hongsheng Li. Pv-rcnn++: Point-voxel feature set abstraction with local vector representa- tion for 3d object detection. Int. J. Comput. Vis., 131(2):531–551, 2023

    2023

  17. [25]

    Pointrcnn: 3d object proposal generation and detection from point cloud

    Shaoshuai Shi, Xiaogang Wang, and Hongsheng Li. Pointrcnn: 3d object proposal generation and detection from point cloud. In IEEE Conf. Comput. Vis. Pattern Recog., pages 770–779, 2019

    2019

  18. [26]

    Semantic scene completion from a single depth image

    Shuran Song, Fisher Yu, Andy Zeng, Angel X Chang, Manolis Savva, and Thomas Funkhouser. Semantic scene completion from a single depth image. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1746–1754, 2017

    2017

  19. [27]

    Scalability in perception for autonomous driving: Waymo open dataset

    Pei Sun, Henrik Kretzschmar, Xerxes Dotiwalla, Aurelien Chouard, Vijaysai Patnaik, Paul Tsui, James Guo, Yin Zhou, Yuning Chai, Benjamin Caine, et al. Scalability in perception for autonomous driving: Waymo open dataset. In IEEE Conf. Comput. Vis. Pattern Recog., pages 2446–2454, 2020

    2020

  20. [28]

    Swformer: Sparse window transformer for 3d object detection in point clouds

    Pei Sun, Mingxing Tan, Weiyue Wang, Chenxi Liu, Fei Xia, Zhaoqi Leng, and Dragomir Anguelov. Swformer: Sparse window transformer for 3d object detection in point clouds. In Eur. Conf. Comput. Vis., pages 426–442. Springer, 2022

    2022

  21. [29]

    Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving

    Xiaoyu Tian, Tao Jiang, Longfei Yun, Yucheng Mao, Huitong Yang, Yue Wang, Yilun Wang, and Hang Zhao. Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving. Adv. Neural Inform. Process. Syst., 36, 2024

    2024

  22. [30]

    Scene as occupancy

    Wenwen Tong, Chonghao Sima, Tai Wang, Li Chen, Silei Wu, Hanming Deng, Yi Gu, Lewei Lu, Ping Luo, Dahua Lin, et al. Scene as occupancy. In Int. Conf. Comput. Vis., pages 8406–8415, 2023

    2023

  23. [31]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. In Adv. Neural Inform. Process. Syst., pages 5998–6008, 2017

    2017

  24. [32]

    Immortal tracker: Tracklet never dies

    Qitai Wang, Yuntao Chen, Ziqi Pang, Naiyan Wang, and Zhaoxiang Zhang. Immortal tracker: Tracklet never dies. arXiv preprint arXiv:2111.13672, 2021. 12

    2021 arXiv

  25. [33]

    Openoccupancy: A large scale benchmark for surrounding semantic occupancy perception

    Xiaofeng Wang, Zheng Zhu, Wenbo Xu, Yunpeng Zhang, Yi Wei, Xu Chi, Yun Ye, Dalong Du, Jiwen Lu, and Xingang Wang. Openoccupancy: A large scale benchmark for surrounding semantic occupancy perception. arXiv preprint arXiv:2303.03991, 2023

    2023 arXiv

  26. [34]

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

    Yi Wei, Linqing Zhao, Wenzhao Zheng, Zheng Zhu, Jie Zhou, and Jiwen Lu. Surroundocc: Multi-camera 3d occupancy prediction for autonomous driving. In Int. Conf. Comput. Vis., pages 21729–21740, 2023

    2023

  27. [35]

    Scpnet: Semantic scene completion on point cloud

    Zhaoyang Xia, Youquan Liu, Xin Li, Xinge Zhu, Yuexin Ma, Yikang Li, Yuenan Hou, and Yu Qiao. Scpnet: Semantic scene completion on point cloud. In IEEE Conf. Comput. Vis. Pattern Recog., pages 17642–17651, 2023

    2023

  28. [36]

    Sparse single sweep lidar point cloud segmentation via learning contextual shape priors from scene completion

    Xu Yan, Jiantao Gao, Jie Li, Ruimao Zhang, Zhen Li, Rui Huang, and Shuguang Cui. Sparse single sweep lidar point cloud segmentation via learning contextual shape priors from scene completion. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 31...

    2021

  29. [37]

    3d-man: 3d multi-frame attention network for object detection

    Zetong Yang, Yin Zhou, Zhifeng Chen, and Jiquan Ngiam. 3d-man: 3d multi-frame attention network for object detection. In IEEE Conf. Comput. Vis. Pattern Recog., pages 1863–1872, 2021

    2021

  30. [38]

    Center-based 3d object detection and tracking

    Tianwei Yin, Xingyi Zhou, and Philipp Krähenbühl. Center-based 3d object detection and tracking. IEEE Conf. Comput. Vis. Pattern Recog., 2021

    2021

  31. [39]

    Monosdf: Exploring monocular geometric cues for neural implicit surface reconstruction

    Zehao Yu, Songyou Peng, Michael Niemeyer, Torsten Sattler, and Andreas Geiger. Monosdf: Exploring monocular geometric cues for neural implicit surface reconstruction. Adv. Neural Inform. Process. Syst., 35:25018–25032, 2022

    2022

  32. [40]

    Hednet: A hierarchical encoder-decoder network for 3d object detection in point clouds

    Gang Zhang, Chen Junnan, Guohuan Gao, Jianmin Li, and Xiaolin Hu. Hednet: A hierarchical encoder-decoder network for 3d object detection in point clouds. Adv. Neural Inform. Process. Syst., 36, 2024

    2024

  33. [41]

    Centerformer: Center-based transformer for 3d object detection

    Zixiang Zhou, Xiangchen Zhao, Yu Wang, Panqu Wang, and Hassan Foroosh. Centerformer: Center-based transformer for 3d object detection. In Eur. Conf. Comput. Vis., pages 496–513. Springer, 2022. 13 Free UnobservedOccupied GT Object SequenceDensified Object PointsObject V olumeO...

    2022

Pith tools

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