Pith. sign in

REVIEW 2 major objections 5 minor 61 references

Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset

T0 review · 2 major / 5 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 MVP-Occ is a synthetic multi-view dataset with voxel-level pedestrian occupancy and panoptic labels, and OmniOcc predicts both 2D and 3D occupancy from camera images, transferring to real surveillance footage.

desk verdict MVP-Occ is a genuinely useful new dataset for dense multi-view pedestrian occupancy; the OmniOcc baseline is competent, but the real-world 3D numbers are 2D proxy scores, so treat that claim with caution. read the letter →

arxiv 2412.13569 v1 pith:IUYZBXYW submitted 2024-12-18 cs.CV

Sithu Aung , Min-Cheol Sagong , Junghyun Cho This is my paper · ORCID
classification cs.CV
keywords multi-viewpedestriandetectionoccupancypredictionvoxelpanopticsegmentationsyntheticdatasetsynthetic-to-realtransfersemanticsceneunderstandingbenchmark
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 new task target for multi-view pedestrian understanding: instead of only predicting locations on the ground plane, predict the occupancy state of every voxel in a large scene, together with semantic, instance, and panoptic labels. To make that possible, it introduces MVP-Occ, a CARLA-generated synthetic dataset of five dense urban scenes with 10 cm voxel annotations (Free, Pedestrian, Ground, Wall, Others), depth, point clouds, and per-person pose and segmentation labels. The paper also builds OmniOcc, a baseline that projects multi-view image features into a voxel grid, processes them with a 3D U-Net, and outputs 2D ground-plane occupancy plus 3D semantic, instance, and panoptic occupancy. On same-scene evaluation OmniOcc reports high voxel IoU and detection scores, and on synthetic-to-real transfer to WildTrack and CityStreet it reports substantially higher F1 than ground-plane-only detectors. The practical stake is a benchmark that lets occupancy prediction and scene understanding be evaluated together in crowded, camera-surveillance-like environments.

What carries the argument

The central mechanism is a non-parametric view transformer that distributes each 2D image feature along the camera ray into a voxel grid, averaging the feature samples from all views that see the voxel; this avoids learned cross-view attention and keeps the model compatible with variable camera counts and scene sizes. A 3D U-Net then refines the voxel feature volume, a two-layer MLP classifies each voxel into Free, Pedestrian, Ground, Wall, or Others, and a separate head averages along the vertical axis and applies one convolution to produce the 2D ground-plane occupancy heatmap. Pedestrian instance grouping assigns each pedestrian-labeled voxel to the nearest detected location within 50 cm, turning the semantic volume into instance and panoptic occupancy. The accompanying weighted cross-entropy, Lovász-Softmax, and scene-class affinity losses balance the dominant Free voxels and sharpen pedestrian boundaries, and the ablation results show that these pieces jointly carry the reported accuracy.

What would settle it

Train OmniOcc on MVP-Occ and evaluate it on a real multi-camera scene that has genuine 3D occupancy ground truth, for example voxelized LiDAR or dense multi-view reconstructions, then compare voxel IoU with the paper's ray-marched mask metric; if the voxel IoU is much lower than the mask-based score, the proxy overstates 3D accuracy.

Watch

Extended reading notes

Core claim

In the paper's own terms, the discovery is that voxel-level pedestrian occupancy can be predicted from multiple posed cameras with high accuracy, and that doing so transfers to real scenes much better than ground-plane localization alone. MVP-Occ is claimed to be the first multi-view pedestrian dataset to provide voxel occupancy and panoptic annotations for large-scale dense scenes: 12,500 frames across five scenes with up to 100 pedestrians per scene and 10 cm voxels. OmniOcc is claimed to outperform the prior multi-view detection baselines MVDet, SHOT, GMVD, and MVFP on the 2D occupancy benchmark, reaching an average F1 of 96.8 in same-scene evaluation, and to exceed them on synthetic-to-real transfer, reaching 87.5 F1 from the Facade scene to WildTrack and 89.7 F1 when trained on all MVP-Occ scenes. The paper further argues that semantic scene understanding, trained through the voxel labels, is a load-bearing component: removing it drops 2D pedestrian F1 from 87.5 to 81.6 and instance AP from 40.2 to 38.2 in the WildTrack transfer setting.

Load-bearing premise

The evaluation of 3D occupancy on real data assumes that agreement between ray-marched 2D segmentation masks and manually annotated masks faithfully measures 3D voxel correctness; if that proxy is not faithful, the reported synthetic-to-real 3D metrics do not establish the 3D occupancy claim.

Editorial extensions

If this is right

  • MVP-Occ gives the multi-view pedestrian community a common benchmark with voxel-level semantic, instance, and panoptic labels, so future methods can be compared on scene understanding and occupancy, not only ground-plane locations.
  • If OmniOcc's transfer results hold, synthetic training with voxel supervision can substitute for expensive real-world 3D labels in dense pedestrian surveillance.
  • The Facade scene, designed to resemble WildTrack, achieves 87.5 F1 in direct transfer, and training on all five scenes raises this to 89.7, implying that scene replication plus diversity is a practical recipe.
  • OmniOcc's simple components—averaging instead of learned compression, a single convolutional head instead of dilated or LKR heads—suggest that extra architectural complexity is not what drives the gains.
  • Because the paper reports 57.3 MODA and 78.0 F1 on CityStreet after training on MVP-Occ, the dataset generalizes beyond its closest WildTrack analog to a larger, noisier real scene.

Reading between the lines

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

  • Editorial inference: the same voxel-occupancy formulation should extend to other dynamic classes such as vehicles or cyclists in multi-camera surveillance, since only the instance-grouping step is pedestrian-specific; a direct test would be annotating a multi-class version of a scene and retraining the baseline.
  • Editorial inference: if voxel supervision is what enables transfer, then densifying real-world labels through LiDAR voxelization or multi-view reconstruction, rather than manual 2D masks, should improve measured 3D accuracy further; the paper does not test this.
  • Editorial inference: the paper's own future-work mention of neural-rendering supervision implies a path where 3D occupancy is learned from 2D labels alone; a testable consequence is that OmniOcc-style models could train without any voxel ground truth, which the current experiments do not isolate.
  • Editorial inference: because the manual 2D-mask proxy may be insensitive to small voxel errors, an independent benchmark with true voxel ground truth would likely rank methods differently than the reported mask-based metrics.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper introduces MVP-Occ, a synthetic multi-view dataset of five CARLA urban scenes with dense pedestrian crowds, providing voxel-level semantic occupancy labels (Free, Pedestrian, Ground, Wall, Others) as well as panoptic, instance, depth, and segmentation annotations. The authors also propose OmniOcc, a baseline model that predicts 2D bird's-eye-view pedestrian occupancy and 3D semantic, instance, and panoptic occupancy from multiple posed images. Experiments report same-scene results on MVP-Occ and synthetic-to-real transfer to WildTrack and CityStreet, and the paper claims state-of-the-art 2D pedestrian occupancy accuracy and strong 3D occupancy transfer performance.

Significance. If the dataset is released, it fills a real gap: existing multi-view pedestrian datasets such as WildTrack, MultiviewX, and GMVD provide only ground-plane locations, while MVP-Occ offers dense voxel-level occupancy and panoptic labels for large outdoor scenes. The dataset construction is described in unusual detail, including the fusion of monocular depth point clouds, the use of additional cameras to handle occlusion, and the labeling protocol. The OmniOcc baseline is simple and well ablated, and the authors provide multiple ablation tables (loss functions, semantic scene understanding, pedestrian instance grouping, location decoder, voxel encoder, backbone). The inclusion of external real-data tests on WildTrack and CityStreet is a positive feature that goes beyond self-evaluation. However, the paper's central claims of synthetic-to-real 3D occupancy performance and of superiority over prior detectors are weakened by two evaluation issues: the real-world 3D metrics are actually 2D segmentation proxies after ray marching, and the comparisons against prior methods are not controlled for the additional voxel-level supervision available to OmniOcc.

major comments (2)
  1. [Supplementary, 'Details about Synthetic-to-Real Evaluation'; Table 6]
  2. [Tables 3 and 5; 'Same-Scene Evaluation' and 'Synthetic-to-Real Evaluation']
minor comments (5)
  1. [References]
  2. [Supplementary, 'Training on all scenes']
  3. [Table 3 footnote]
  4. [Table 6 caption]
  5. [Supplementary, 'Details about Synthetic-to-Real Evaluation']

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: external real-data benchmarks and independent baselines support the claims; the 2D segmentation proxy in Table 6 is a stated limitation, not a self-referential reduction.

full rationale

The paper's central claims are a new synthetic dataset (MVP-Occ) and a baseline model (OmniOcc) for multi-view pedestrian occupancy prediction. The same-scene evaluation uses held-out frames of the authors' own synthetic dataset, which is a standard benchmark-construction practice rather than a circular derivation. The synthetic-to-real evaluation is conducted on WildTrack and CityStreet, external real datasets, with comparisons against MVDet, SHOT, GMVD, and MVFP; MVFP is the authors' previous work but is used only as a baseline, not as a load-bearing justification for the present model or dataset. No fitted parameter or self-citation defines the target predictions by construction. The one caveat is the supplementary admission that WildTrack lacks 3D occupancy ground truth, so Table 6 'gauges 3D occupancy prediction performance with 2D segmentation labels' via voxel ray marching; this is an explicitly stated evaluation proxy and a correctness limitation, not a circular reduction, because the 2D masks are manually annotated independently of the model and the metrics are adapted to view-level segmentation rather than being defined in terms of the model's own outputs. The non-parametric feature transformation is cited from prior work as an architectural choice, not as an external uniqueness theorem or ansatz that forces the result. Therefore, no circular step can be exhibited from the paper's equations or citations.

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

The central claims rest on standard multi-view geometry, the CARLA simulator's realism, the accuracy of depth-fused occupancy labels, and the validity of the 2D-mask proxy evaluation on real data. No new physical entities are postulated. The listed free parameters are hand-set evaluation and training choices; none are fitted to a target result in the sense of a scientific derivation.

free parameters (5)
  • Voxel size = 0.10 m
    Chosen for discretization; directly sets resolution of occupancy labels and rendered segmentation quality, and affects metric results.
  • Confidence threshold tau = 0.5
    Used to filter 2D occupancy detections before NMS and instance grouping.
  • Grouping radius r = 0.50 m
    Defines the binary mask W in pedestrian instance grouping; determines voxel-to-instance assignment.
  • Loss weights = 0.4, 0.3, 0.3, 0.3
    Hand-set hyperparameters in total loss Eq. (7)-(8); no sensitivity analysis is provided.
  • WildTrack AP IoU thresholds = T = {0.25, 0.30, ..., 0.70}
    Adjusted from COCO T = {0.50, ..., 0.95} because of severe occlusion in 2D views; makes reported AP non-standard.
assumptions (5)
  • standard math Pinhole camera model with known intrinsics K and extrinsics R, used in projection equations (1)-(3), is valid.
    Used to project voxel centers into image feature maps in the View Transformer section.
  • domain assumption Uniform feature distribution along camera rays: every voxel on a ray receives the same image feature as the pixel it projects to.
    View Transformer adopts a non-parametric feature transformation from ImVoxelNet, assuming uniform depth distribution along rays.
  • domain assumption CARLA-simulated pedestrians, scenes, and camera setups are representative enough that models trained on MVP-Occ transfer to real surveillance data.
    Synthetic-to-real experiments on WildTrack and CityStreet rest on this assumption.
  • domain assumption The fused depth-based point clouds from monocular views produce accurate occupancy labels for MVP-Occ.
    Occupancy label generation in the supplementary fuses monocular depth point clouds because CARLA mesh extraction and LiDAR were unavailable.
  • domain assumption Manually annotated 2D semantic, instance, and panoptic masks on WildTrack are reliable proxies for evaluating 3D occupancy predictions.
    Synthetic-to-real evaluation uses these masks rendered from ray-marched occupancy predictions instead of real 3D labels.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset." pith.science (2026). https://pith.science/paper/IUYZBXYW

@misc{pith2026241213569,
  author       = {Pith},
  title        = {Pith review of: Multi-View Pedestrian Occupancy Prediction with a Novel Synthetic Dataset},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IUYZBXYW}},
  note         = {Machine review of arXiv:2412.13569}
}
read the original abstract

We address an advanced challenge of predicting pedestrian occupancy as an extension of multi-view pedestrian detection in urban traffic. To support this, we have created a new synthetic dataset called MVP-Occ, designed for dense pedestrian scenarios in large-scale scenes. Our dataset provides detailed representations of pedestrians using voxel structures, accompanied by rich semantic scene understanding labels, facilitating visual navigation and insights into pedestrian spatial information. Furthermore, we present a robust baseline model, termed OmniOcc, capable of predicting both the voxel occupancy state and panoptic labels for the entire scene from multi-view images. Through in-depth analysis, we identify and evaluate the key elements of our proposed model, highlighting their specific contributions and importance.

Figures

Figures reproduced from arXiv: 2412.13569 by the authors.

Figure 1
Figure 1. Visualizations of the proposed dataset. The pri￾mary objective is to predict the semantic and instance la￾bels of the voxels and determine each pedestrian’s location within the scene. The dataset includes five expansive scenes with dense pedestrian activity. (Best viewed in color.) allows us to identify each pedestrian instance, while voxel￾level occupancy prediction provides rich geometric informa￾tion about the sc… view at source ↗
Figure 2
Figure 2. Overview of the proposed model. Image features are extracted using a backbone network augmented with an FPN. Next, multi-view 2D features are projected onto the voxel grid along rays and processed with a 3D U-Net to construct a feature volume. Semantic occupancy predictions are generated using a two-layer MLP network, whereas a single convolutional layer predicts the occupancy status of pedestrians. Finally, pedestr… view at source ↗
Figure 3
Figure 3. Qualitative results of 2D and 3D occupancy pre￾dictions under same-scene evaluation on the Park scene. (Best viewed in color.) Experiments In this section, we first report on the evaluation metrics and training details. The subsequent sections provide exten￾sive experiments on both 2D and 3D occupancy prediction benchmarks and lastly, ablation studies. Evaluation metrics. We use commonly adopted metrics from multi-v… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Qualitative results of synthetic-to-real transfer from Facade to WildTrack. (Best viewed in color.) mean IoU across all classes. We report AP (Average Preci￾sion) as mask AP using pycocotools (Lin et al. 2014) for 3D instance occupancy and PQ (Panoptic Quality) metric …
Figure 5
Figure 5. Figure 5: Camera views used for generating a scene point cloud. The number of additional camera views varies depending on the scene’s characteristics, such as its size and level of occlusion. However, every scene includes an overhead camera to ensure comprehensive coverage. 3D p…
Figure 6
Figure 6. Figure 6: Comparison between ground-truth and rendered segmentation masks. Multi-view consistent rendered segmen￾tation data are generated from our model trained on the Facade scene (third and last rows). Location Decoder 2D Ped. Occ. 3D Sem. Occ. 3D Pan. Occ. Compress Head MODA…
Figure 7
Figure 7. Figure 7: Qualitative results of 3D panoptic occupancy prediction under synthetic-to-real evaluation on WildTrack. The model is trained on each scene of MVP-Occ and tested on the WildTrack. Note that there is no ground-truth 3D occupancy data available in the WildTrack scene [P…
Figure 8
Figure 8. Figure 8: Qualitative results of 3D panoptic occupancy prediction under same-scene evaluation on MVP-Occ. The model is trained and tested on each scene but with different splits [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

61 extracted references · 42 canonical work pages

  1. [1]

    Amanatides, J.; Woo, A.; et al. 1987. A fast voxel traversal algorithm for ray tracing. In Eurographics, volume 87, 3--10. Citeseer

    work page 1987

  2. [2]

    Aung, S.; Park, H.; Jung, H.; and Cho, J. 2024. Enhancing Multi-View Pedestrian Detection Through Generalized 3D Feature Pulling. In WACV, 1196--1205

    work page 2024

  3. [3]

    R.; and Blaschko, M

    Berman, M.; Triki, A. R.; and Blaschko, M. B. 2018. The lov \'a sz-softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks. In CVPR, 4413--4421

    work page 2018

  4. [4]

    Cao, A.-Q.; and de Charette, R. 2022. Monoscene: Monocular 3d semantic scene completion. In CVPR, 3991--4001

    work page 2022

  5. [5]

    Chavdarova, T.; Baqu \'e , P.; Bouquet, S.; Maksai, A.; Jose, C.; Bagautdinov, T.; Lettry, L.; Fua, P.; Van Gool, L.; and Fleuret, F. 2018. Wildtrack: A multi-camera hd dataset for dense unscripted pedestrian detection. In CVPR, 5030--5039

    work page 2018

  6. [6]

    Chen, X.; Lin, K.-Y.; Qian, C.; Zeng, G.; and Li, H. 2020. 3d sketch-aware semantic scene completion via semi-supervised structure prior. In CVPR, 4193--4202

    work page 2020

  7. [7]

    S.; Brox, T.; and Ronneberger, O

    C i c ek, \"O .; Abdulkadir, A.; Lienkamp, S. S.; Brox, T.; and Ronneberger, O. 2016. 3D U-Net: learning dense volumetric segmentation from sparse annotation. In Medical Image Computing and Computer-Assisted Intervention--MICCAI 2016: 19th International Conference, Athens, Greece, October 17-21, 2016, Proceedings, Part II 19, 424--432. Springer

    work page 2016

  8. [8]

    Dong, J.; Jiang, W.; Huang, Q.; Bao, H.; and Zhou, X. 2019. Fast and robust multi-person 3d pose estimation from multiple views. In CVPR, 7792--7801

    work page 2019

Show all 61 references
  1. [9]

    Dosovitskiy, A.; Ros, G.; Codevilla, F.; Lopez, A.; and Koltun, V. 2017. CARLA: An open urban driving simulator. In Conference on robot learning, 1--16. PMLR

    2017

  2. [10]

    Fan, B.; Wang, S.; Zheng, W.; Feng, J.; and Zhou, J. 2023. Human-M3: A Multi-view Multi-modal Dataset for 3D Human Pose Estimation in Outdoor Scenes. arXiv preprint arXiv:2308.00628

    2023 arXiv

  3. [11]

    Fang, M.; and Li, Z. 2023. Occupancy Dataset for nuScenes . https://github.com/FANG-MING/occupancy-for-nuscenes

    2023

  4. [12]

    Gan, W.; Mo, N.; Xu, H.; and Yokoya, N. 2023. A Simple Attempt for 3D Occupancy Estimation in Autonomous Driving. arXiv preprint arXiv:2303.10076

    2023 arXiv

  5. [13]

    Gan, Y.; Han, R.; Yin, L.; Feng, W.; and Wang, S. 2021. Self-supervised multi-view multi-human association and tracking. In Proceedings of the 29th ACM International Conference on Multimedia, 282--290

    2021

  6. [14]

    W.; Fang, Z.; Li, J.; Ambrus, R.; and Fragkiadaki, K

    Harley, A. W.; Fang, Z.; Li, J.; Ambrus, R.; and Fragkiadaki, K. 2023. Simple-bev: What really matters for multi-sensor bev perception? In ICRA, 2759--2765. IEEE

    2023

  7. [15]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In CVPR, 770--778

    2016

  8. [16]

    Hou, Y.; and Zheng, L. 2021. Multiview detection with shadow transformer (and view-coherent data augmentation). In Proceedings of the 29th ACM International Conference on Multimedia, 1673--1682

    2021

  9. [17]

    Hou, Y.; Zheng, L.; and Gould, S. 2020. Multiview Detection with Feature Perspective Transformation. In ECCV

    2020

  10. [18]

    Huang, Y.; Zheng, W.; Zhang, B.; Zhou, J.; and Lu, J. 2023 a . SelfOcc: Self-Supervised Vision-Based 3D Occupancy Prediction. arXiv preprint arXiv:2311.12754

    2023 arXiv

  11. [19]

    Huang, Y.; Zheng, W.; Zhang, Y.; Zhou, J.; and Lu, J. 2023 b . Tri-perspective view for vision-based 3d semantic occupancy prediction. In CVPR, 9223--9232

    2023

  12. [20]

    Q.; Park, M.; and Park, S

    Jeon, Y.; Tran, D. Q.; Park, M.; and Park, S. 2023. Leveraging Future Trajectory Prediction for Multi-Camera People Tracking. In CVPR, 5398--5407

    2023

  13. [21]

    Joo, H.; Liu, H.; Tan, L.; Gui, L.; Nabbe, B.; Matthews, I.; Kanade, T.; Nobuhara, S.; and Sheikh, Y. 2015. Panoptic Studio: A Massively Multiview System for Social Motion Capture. In ICCV

    2015

  14. [22]

    Kirillov, A.; He, K.; Girshick, R.; Rother, C.; and Doll \'a r, P. 2019. Panoptic segmentation. In CVPR, 9404--9413

    2019

  15. [23]

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

    2023 arXiv

  16. [24]

    Lin, J.; and Lee, G. H. 2021. Multi-view multi-person 3d pose estimation with plane sweep stereo. In CVPR, 11886--11895

    2021

  17. [25]

    Lin, T.-Y.; Doll \'a r, P.; Girshick, R.; He, K.; Hariharan, B.; and Belongie, S. 2017. Feature pyramid networks for object detection. In CVPR, 2117--2125

    2017

  18. [26]

    Lin, T.-Y.; Maire, M.; Belongie, S.; Hays, J.; Perona, P.; Ramanan, D.; Doll \'a r, P.; and Zitnick, C. L. 2014. Microsoft coco: Common objects in context. In ECCV, 740--755. Springer

    2014

  19. [27]

    Liu, H.; Wang, H.; Chen, Y.; Yang, Z.; Zeng, J.; Chen, L.; and Wang, L. 2023. Fully Sparse 3D Panoptic Occupancy Prediction. arXiv preprint arXiv:2312.17118

    2023 arXiv

  20. [28]

    C.; Mart \' nez, J

    Luna, E.; SanMiguel, J. C.; Mart \' nez, J. M.; and Carballeira, P. 2022. Graph Neural Networks for Cross-Camera Data Association. IEEE Transactions on Circuits and Systems for Video Technology, 33(2): 589--601

    2022

  21. [29]

    Miao, R.; Liu, W.; Chen, M.; Gong, Z.; Xu, W.; Hu, C.; and Zhou, S. 2023. Occdepth: A depth-aware method for 3d semantic scene completion. arXiv preprint arXiv:2302.13540

    2023 arXiv

  22. [30]

    P.; Tancik, M.; Barron, J

    Mildenhall, B.; Srinivasan, P. P.; Tancik, M.; Barron, J. T.; Ramamoorthi, R.; and Ng, R. 2021. Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65(1): 99--106

    2021

  23. [31]

    Milioto, A.; Vizzo, I.; Behley, J.; and Stachniss, C. 2019. Rangenet++: Fast and accurate lidar semantic segmentation. In 2019 IEEE/RSJ international conference on intelligent robots and systems (IROS), 4213--4220. IEEE

    2019

  24. [32]

    Murez, Z.; Van As, T.; Bartolozzi, J.; Sinha, A.; Badrinarayanan, V.; and Rabinovich, A. 2020. Atlas: End-to-end 3d scene reconstruction from posed images. In ECCV, 414--431. Springer

    2020

  25. [33]

    Q.-V.; Le, H

    Nguyen, Q. Q.-V.; Le, H. D.-A.; Chau, T. T.-T.; Luu, D. T.; Chung, N. M.; and Ha, S. V.-U. 2023. Multi-camera people tracking with mixture of realistic and synthetic knowledge. In CVPR, 5495--5505

    2023

  26. [34]

    Pan, M.; Liu, J.; Zhang, R.; Huang, P.; Li, X.; Liu, L.; and Zhang, S. 2023. RenderOcc: Vision-Centric 3D Occupancy Prediction with 2D Rendering Supervision. arXiv preprint arXiv:2309.09502

    2023 arXiv

  27. [35]

    S.; and Yang, X

    Qiu, R.; Xu, M.; Yan, Y.; Smith, J. S.; and Yang, X. 2022. 3d random occlusion and multi-layer projection for deep multi-camera pedestrian localization. In ECCV, 695--710. Springer

    2022

  28. [36]

    Roldao, L.; de Charette, R.; and Verroust-Blondet, A. 2020. Lmscnet: Lightweight multiscale 3d semantic completion. In 3DV, 111--119. IEEE

    2020

  29. [37]

    Rukhovich, D.; Vorontsova, A.; and Konushin, A. 2022. Imvoxelnet: Image to voxels projection for monocular and multi-view general-purpose 3d object detection. In WACV, 2397--2406

    2022

  30. [38]

    Sima, C.; Renz, K.; Chitta, K.; Chen, L.; Zhang, H.; Xie, C.; Luo, P.; Geiger, A.; and Li, H. 2023. Drivelm: Driving with graph visual question answering. arXiv preprint arXiv:2312.14150

    2023 arXiv

  31. [39]

    Song, L.; Wu, J.; Yang, M.; Zhang, Q.; Li, Y.; and Yuan, J. 2021. Stacked homography transformations for multi-view pedestrian detection. In ICCV, 6049--6057

    2021

  32. [40]

    X.; Savva, M.; and Funkhouser, T

    Song, S.; Yu, F.; Zeng, A.; Chang, A. X.; Savva, M.; and Funkhouser, T. 2017. Semantic scene completion from a single depth image. In CVPR, 1746--1754

    2017

  33. [41]

    Sun, T.; Segu, M.; Postels, J.; Wang, Y.; Van Gool, L.; Schiele, B.; Tombari, F.; and Yu, F. 2022. SHIFT: A Synthetic Driving Dataset for Continuous Multi-Task Domain Adaptation. In CVPR, 21371--21382

    2022

  34. [42]

    Tang, J.; Chen, X.; Wang, J.; and Zeng, G. 2022. Not all voxels are equal: Semantic scene completion from the point-voxel perspective. In AAAI, volume 36, 2352--2360

    2022

  35. [43]

    Tesla. 2022. Tesla AI Day . https://www.youtube.com/watch?v=ODSJsviD_SU

    2022

  36. [44]

    Tian, X.; Jiang, T.; Yun, L.; Wang, Y.; Wang, Y.; and Zhao, H. 2023. Occ3d: A large-scale 3d occupancy prediction benchmark for autonomous driving. arXiv preprint arXiv:2304.14365

    2023 arXiv

  37. [45]

    Tong, W.; Sima, C.; Wang, T.; Chen, L.; Wu, S.; Deng, H.; Gu, Y.; Lu, L.; Luo, P.; Lin, D.; et al. 2023 a . Scene as occupancy. In ICCV, 8406--8415

    2023

  38. [46]

    Tong, W.; Sima, C.; Wang, T.; Chen, L.; Wu, S.; Deng, H.; Gu, Y.; Lu, L.; Luo, P.; Lin, D.; et al. 2023 b . Scene as occupancy. In ICCV, 8406--8415

    2023

  39. [47]

    Tu, H.; Wang, C.; and Zeng, W. 2020. VoxelPose: Towards Multi-Camera 3D Human Pose Estimation in Wild Environment. In ECCV

    2020

  40. [48]

    Vora, J.; Dutta, S.; Jain, K.; Karthik, S.; and Gandhi, V. 2023. Bringing Generalization to Deep Multi-View Pedestrian Detection. In WACV, 110--119

    2023

  41. [49]

    Wei, Y.; Zhao, L.; Zheng, W.; Zhu, Z.; Zhou, J.; and Lu, J. 2023. Surroundocc: Multi-camera 3d occupancy prediction for autonomous driving. In ICCV, 21729--21740

    2023

  42. [50]

    Weng, X.; Man, Y.; Park, J.; Yuan, Y.; Cheng, D.; O'Toole, M.; and Kitani, K. 2021. All-In-One Drive: A Large-Scale Comprehensive Perception Dataset with High-Density Long-Range Point Clouds . arXiv

    2021

  43. [51]

    Wu, S.-C.; Tateno, K.; Navab, N.; and Tombari, F. 2020. Scfusion: Real-time incremental scene reconstruction with semantic completion. In 3DV, 801--810. IEEE

    2020

  44. [52]

    Yang, J.; Desai, K.; Packer, C.; Bhatia, H.; Rhinehart, N.; McAllister, R.; and Gonzalez, J. 2024. CARFF: Conditional Auto-encoded Radiance Field for 3D Scene Forecasting. arXiv preprint arXiv:2401.18075

    2024 arXiv

  45. [53]

    Yao, J.; Li, C.; Sun, K.; Cai, Y.; Li, H.; Ouyang, W.; and Li, H. 2023. Ndc-scene: Boost monocular 3d semantic scene completion in normalized device coordinates space. In ICCV, 9421--9431. IEEE Computer Society

    2023

  46. [54]

    Zhang, C.; Yan, J.; Wei, Y.; Li, J.; Liu, L.; Tang, Y.; Duan, Y.; and Lu, J. 2023. OccNeRF: Self-Supervised Multi-Camera Occupancy Prediction with Neural Radiance Fields. arXiv preprint arXiv:2312.09243

    2023 arXiv

  47. [55]

    Zhang, J.; Cai, Y.; Yan, S.; Feng, J.; et al. 2021. Direct multi-view multi-person 3d pose estimation. NeurIPS, 34: 13153--13164

    2021

  48. [56]

    Zhang, Q.; and Chan, A. B. 2019 a . Wide-area crowd counting via ground-plane density maps and multi-view fusion cnns. In CVPR, 8297--8306

    2019

  49. [57]

    Zhang, Q.; and Chan, A. B. 2019 b . Wide-area crowd counting via ground-plane density maps and multi-view fusion cnns. In CVPR, 8297--8306

    2019

  50. [58]

    Zhang, Q.; Lin, W.; and Chan, A. B. 2021. Cross-view cross-scene multi-view crowd counting. In CVPR, 557--567

    2021

  51. [59]

    Zhang, Y.; Zhu, Z.; and Du, D. 2023. OccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Prediction. arXiv preprint arXiv:2304.05316

    2023 arXiv

  52. [60]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all...

  53. [61]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

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