Pith. sign in

REVIEW 5 major objections 7 minor 36 references

3D-CovDiffusion: 3D-Aware Diffusion Policy for Coverage Path Planning

T0 review · 5 major / 7 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A single diffusion policy can generate ordered, smooth, high-coverage 6-DoF spray-painting trajectories from raw point clouds across different object categories, the paper argues.

desk verdict The architecture is a reasonable engineering contribution, but the abstract's numbers don't survive contact with the paper's own Table I, and the unmatched trajectory-history conditioning makes the main comparison unfair. read the letter →

arxiv 2510.03011 v2 pith:MDHPFZVV submitted 2025-10-03 cs.RO

Chenyuan Chen , Haoran Ding , Ran Ding , Tianyu Liu , Zewen He , Anqing Duan , Yoshihiko Nakamura This is my paper
classification cs.RO
keywords coveragepathplanningdiffusionpolicy6-DoFtrajectorygenerationpointcloudconditioningspraypaintingimitationlearningroboticsurfaceprocessing
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 coverage path planning for industrial surface processing can be reformulated as conditional sequence generation and solved with one geometry-conditioned diffusion policy. Instead of predicting unordered stroke segments and stitching them with heuristics, the model denoises a full trajectory chunk conditioned on a point cloud and recent motion, producing ordered chunks that concatenate directly. If correct, this removes category-specific training and heuristic post-processing, letting a single policy serve cuboids, windows, shelves, and containers. The reported gains are large—about 98% lower point-wise Chamfer distance and jerk, and much higher surface coverage than prior learning-based baselines. The paper's own ablation, however, shows the model is trained with a ground-truth trajectory-history input that no baseline receives, which complicates the fairness of the headline comparison.

What carries the argument

A conditional denoising diffusion implicit model (DDIM) with a cosine noise schedule, conditioned by a fused 128-dimensional vector. The condition concatenates (a) a global point-cloud descriptor from a three-layer MLP with max pooling (the 3DCovDiff encoder) and (b) a 64-dimensional embedding of the last four 6-DoF poses from a two-layer MLP. The condition is injected into every denoising residual block via FiLM (feature-wise linear modulation), and the training loss masks noise prediction over variable-length trajectories. This mechanism carries the whole claim: iterative denoising produces smooth, ordered, variable-length chunks, while the geometry encoder is what lets one policy generali

What would settle it

Retrain PaintNet or the point-wise baseline with the same four-step ground-truth history as an additional input, or retrain the proposed model without the history channel, on the same train/test split; if the Chamfer-distance gap shrinks from roughly 98% to near parity, the central claim of a superior diffusion-based coverage policy is falsified. Additionally, recompute coverage with spray radius r=0.02 m and r=0.10 m to test whether the coverage gain persists.

Watch

Extended reading notes

Core claim

The central claim is that coverage path planning can be cast as conditional sequence generation and solved by a geometry-conditioned denoising diffusion implicit model (DDIM). The policy consumes a partial point cloud of the target surface plus the most recent four end-effector poses and outputs the next trajectory chunk in temporally ordered form; chunks are concatenated sequentially without heuristic sorting or overlapping-endpoint alignment. Trained once on demonstrations for cuboids, windows, shelves, and containers, the same policy generalizes across those categories, and the paper reports large improvements over PaintNet, multi-path prediction, and point-wise baselines on Chamfer dista

Load-bearing premise

The load-bearing premise is that the comparison with baselines is fair, yet the proposed model is trained with a ground-truth trajectory-history input that no baseline receives, and an unevaluated re-processing of the dataset plus a hand-set spray radius in the coverage metric further affect the reported margins.

Editorial extensions

If this is right

  • Industrial coating, painting, and polishing could use one learned policy per task instead of per object geometry, reducing manual redesign when the part changes.
  • Temporally ordered trajectory chunks can be concatenated sequentially, removing heuristic stitching and overlapping-endpoint alignment from the pipeline.
  • Because the policy outputs a distribution over trajectories, it can represent multiple valid coverage patterns for the same geometry, matching the multimodality of expert demonstrations.
  • The conditioning on recent poses makes the policy auto-regressive, so longer trajectories can be generated by feeding predictions back as context.
  • The 88-sample container category shows data scarcity, not model capacity, is the main brake on generalization to complex geometries.

Reading between the lines

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

  • The 98% Chamfer-distance and jerk margins over PaintNet likely overstate the diffusion contribution: the proposed model alone receives the last four ground-truth poses as input at train time, while no baseline gets any trajectory-history channel; a fair comparison would give the same conditioning to the baselines or remove it from the proposed model.
  • Only the 3D positions are scored by PCD and coverage; orientation is part of the 6-DoF output, so orientation accuracy is unverified. A test that separates position and orientation error could reorder the comparisons.
  • The coverage metric counts a face covered only if a segment passes within the hand-set spray radius (0.05 m) of its centroid; sweeping this radius would show whether the coverage advantage is robust or an artifact of the threshold.
  • A natural testable extension: condition the same diffusion policy on a short history of the model's own predictions instead of ground truth at training time, or evaluate closed-loop generation, to see whether the ordering and smoothness benefits survive without oracle context.
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

5 major / 7 minor

Summary. The paper proposes 3D-CovDiffusion, a DDIM-based conditional diffusion policy that generates ordered 6-DoF spray-painting trajectory chunks from a partial point cloud and a 4-step motion-history condition. The authors claim that a single policy, without category-specific training, can generalize across cuboids, windows, shelves, and containers, and that it outperforms PaintNet, Multi-Path Prediction, and Point-Wise Prediction by large margins in point-wise Chamfer distance, surface coverage, and jerk-based smoothness. The evaluation uses an extended PaintNet/MaskPlanner dataset with 80/20 splits. The paper includes ablations on the point-cloud encoder and on the trajectory-conditioning mechanism.

Significance. The problem is relevant: replacing heuristic segment stitching with a learned, geometry-conditioned sequence generator would be a practical advance for industrial surface processing. If the claimed gains were real, the work would provide a useful baseline for diffusion-based coverage path planning. The paper also has some positive features: it reports mean±std over seeds, runs ablations of encoder choice and trajectory conditioning, and provides qualitative comparisons. However, the quantitative claims are not supported by the paper's own table, and the main comparison is confounded by an input channel (the 4-step history) that baselines do not receive. These are internal problems requiring substantive rework, not presentation fixes.

major comments (5)
  1. [Abstract; Table I; Sec. IV.A.b] The abstract's headline numbers are not derivable from Table I. Coverage: the average gain over PaintNet is +58.2 pp, not +67.5 pp, and vs Point-Wise it is -11.4 pp (Ours is lower on Cuboids, Shelves, Containers). PCD: the average reduction vs PaintNet is 93.2%, not 98.2%; vs Point-Wise there is no positive average because Containers is 2.0x worse (622.16 vs 313.98). Jerk: the average reduction vs PaintNet is 85.7%; the 97.8% figure comes only from the selective comparison with Point-Wise, and even that is not 97.0%. Section IV.A.b acknowledges only the Windows gain over Point-Wise and omits the coverage losses. Since the abstract's central claim rests on these numbers, the headline result is internally inconsistent with the only quantitative comparison.
  2. [Sec. III.C.a, Eqs. (3)-(4); Table III] The proposed model is conditioned on a 4-step, 24-DoF trajectory-history vector; the baselines are not. The ablation in Table III shows that this channel carries nearly the entire PCD gain on Windows: Previous Traj. gives 10.41, No Traj. gives 246.07, the latter being worse than the Point-Wise baseline's 55.71 (Table I). The paper does not state whether the history used at evaluation is the ground-truth prefix or model-generated. If it is the ground-truth prefix, the comparison is teacher-forced and not equivalent to the baselines; if it is model-generated, the protocol must still be applied to baselines. Without baselines that receive an equivalent conditioning signal, the margins in Table I cannot be attributed to the diffusion model or geometry encoder.
  3. [Sec. III.B, Eq. (8); Introduction] A stated contribution is avoiding 'post-hoc heuristic ordering or stitching' via 'simple sequential concatenation.' However, Sec. III.B says 'Segments are aligned by matching end and start poses, and concatenated into the final trajectory.' Matching endpoints is a post-hoc alignment heuristic, similar to PaintNet's overlapping-endpoint stitching. The manuscript should clarify what matching involves and whether this heuristic is used in the reported numbers; as written, the central distinction from the baseline is contradicted.
  4. [Sec. IV.b, Eqs. (10)-(13)] The coverage metric depends on the hand-set spray radius r_spray=0.05 m. No sensitivity analysis is reported, and with a single radius the metric may not discriminate among methods or could reverse the ranking at other plausible radii. Additionally, although the outputs are claimed to be 6-DoF, PCD and coverage use only 3D positions; orientation is never evaluated. A spray-coverage metric for 6-DoF trajectories should at least report whether the orientation is consistent with the surface normal, otherwise the '6-DoF' claim is unverified.
  5. [Sec. III.C.b; Sec. IV.A] The paper claims a single policy generalizes across categories without category-specific training, but the experimental section does not state whether Table I comes from one joint model or four per-category models. If separate models were trained for each category, the generalization claim is not tested. In addition, Sec. III.C.b mentions a 'modified data processing pipeline' with category-specific scaling; it is not stated whether the baseline numbers were recomputed under this pipeline or taken from the original publications. Both points need to be disclosed before the comparison can be interpreted.
minor comments (7)
  1. [Abstract] The two abstracts report different coverage numbers: '+67.5 percentage points' and 'increasing surface coverage by 61%'. Please make the claim consistent.
  2. [Sec. IV.A] The opening calls PCD 'our primary evaluation metric,' while Sec. IV.b calls coverage 'a primary evaluation metric.' Clarify which is primary.
  3. [Sec. IV.b] Smoothness is described as 'jerk statistics' but no definition, equation, or units are given.
  4. [Sec. III.C.a; Table III] The text mentions a 'Random Traj.' variant with PCD 264.89, but Table III lists only Zero Traj., No Traj., and Previous Traj. Add the missing row or remove the mention.
  5. [Fig. 4] The caption says 'Columns (left to right) show PaintNet, and 3DCovDiffusion (Ours),' but the figure header includes Multi-path prediction as a middle column. The caption is inconsistent with the figure.
  6. [Sec. I; end of paper] The code link is a placeholder ('can be found here') with no URL. Please provide the actual link or state that code will be released.
  7. [Table II] The text quotes PointNet area-weighted coverage as 98.85%, while Table II gives 98.83%. Align the numbers.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found; headline-vs-Table I discrepancies and unmatched trajectory conditioning are reporting/benchmark issues, not circularity.

full rationale

The derivation chain is self-contained in the sense required for circularity analysis. Equations (1)-(8) define a standard DDIM denoiser conditioned on c=concat(f_pc(P), f_traj(τ̂)) (Eq. 3), trained to predict added noise (Eqs. 2, 6) and sampled via Eq. 7. Nothing in these equations defines the predicted trajectory as the conditioning input itself; at inference τ̂ is the model's own prior output, and the target τ is a separate future trajectory. The large conditioning effect in Table III (Previous Traj. PCD 10.41 vs No Traj. 246.07 / Zero Traj. 284.46) is an empirical ablation result, not an identity or construction. It indicates a possible benchmark-fairness confound because the PaintNet/Point-Wise baselines do not receive the trajectory-history channel, but that is an experimental-design criticism, not the kind of definitional circularity defined in the rubric. Encoder/citation dependencies are external: DP3 [3], PaintNet [2], and MaskPlanner [5] are not self-citations, no uniqueness theorem is imported from the authors, and no ansatz is smuggled via a self-citation. The paper's own numbers are internally inconsistent — the abstract's 98.2% PCD reduction vs PaintNet is not the 93.2% average from Table I, coverage gains vs Point-Wise are negative in three categories (Cuboids, Shelves, Containers), and the advertised +67.5 pp is not recoverable from the table — and Section V.b acknowledges data-scarcity limitations for containers. These are correctness/reproducibility concerns, not circularity; per the hard rules, such issues do not raise the circularity score. Hence score 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The central comparison rests on hand-set evaluation constants (r_spray, trajectory window length, per-category scaling), on an unmatched conditioning input, and on the assumption that GT demonstrations and the line-segment coverage model are faithful proxies for real coating quality. No new physical entities are postulated; the only 'new' object is the renamed point encoder.

free parameters (4)
  • r_spray (coverage spray radius) = 0.05 m (default, never varied)
    The coverage rate (Eqs. 10-12) counts a mesh face as covered iff a trajectory segment passes within r_spray of its centroid. All coverage comparisons, including the headline +67.5pp claim, hinge on this hand-set threshold; no sensitivity analysis or physical calibration is given.
  • Trajectory history length m (conditioning window) = 4 steps = 24 DoF
    Sec. III.C.a. The recent-trajectory conditioning channel is the dominant performance factor (Table III: 10.41 vs 246.07 PCD). The window size is chosen without justification or ablation.
  • Category-specific scaling of point clouds and GT trajectories = per-category scale (values not reported)
    Sec. III.C.b: 'includes category-specific scaling of point clouds and ground-truth trajectories.' Per-category normalization affects PCD and coverage for every method and is not reported, so its role in the comparison cannot be audited.
  • Guidance scale (DDIM/classifier-free) = 2.5
    Sec. III.C.a. Chosen by hand; the paper never states how unconditional training is performed to enable classifier-free guidance, so the sampling procedure (Eq. 7) is incompletely specified.
assumptions (4)
  • domain assumption Expert demonstrations in the PaintNet/MaskPlanner dataset are complete, consistent, and adequate coverage exemplars for each category.
    The whole training/evaluation loop treats demos as ground truth for PCD and coverage; no analysis of demonstration quality or coverage completeness is given (Sec. IV).
  • domain assumption Line-segment distance to face centroids within r_spray is a valid physical model of spray deposition coverage.
    Eqs. 10-12: coverage is defined geometrically with a single radius; deposition physics, paint thickness, and overspray are ignored despite the claims about real-world coating completeness.
  • ad hoc to paper Endpoint-matching alignment of generated segments yields a trajectory as good as the expert's, without heuristics.
    The paper claims end-to-end ordered generation avoids stitching heuristics, yet Eq. 8 aligns segments by 'matching end and start poses', a stitching operation whose failure modes (drift, duplication, gaps) are unexamined.
  • standard math Standard DDIM convergence: training the noise predictor to minimize Eq. 6 and sampling Eq. 7 recovers the demonstration distribution on variable-length 6DoF trajectories.
    Standard diffusion theory from [6,7]; invoked as background.
invented entities (1)
  • 3DCovDiff encoder
    purpose: Geometry conditioning encoder (3-layer MLP, LayerNorm, max pool, projection) claimed as a contribution.
    Architecturally near-identical to the DP3 encoder cited as inspiration [3] (and to PointNet [31]); no ablation isolates the encoder design from the rest of the pipeline except Table II's comparison against other standard encoders, where the PCD advantage (10.41 vs 25.25 for Point Transformer) partly reflects the same confounding.

how reviews work

0 comments
Cite this review

Pith. "Pith review of 3D-CovDiffusion: 3D-Aware Diffusion Policy for Coverage Path Planning." pith.science (2026). https://pith.science/paper/MDHPFZVV

@misc{pith2026251003011,
  author       = {Pith},
  title        = {Pith review of: 3D-CovDiffusion: 3D-Aware Diffusion Policy for Coverage Path Planning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MDHPFZVV}},
  note         = {Machine review of arXiv:2510.03011}
}
read the original abstract

Diffusion models have shown strong potential for robot skill learning, yet their role in coverage path planning remains underexplored. In industrial surface processing (painting, polishing, spray coating), high coverage requires globally ordered, temporally coherent trajectories rather than stitching unordered local segments. We reformulate coverage path planning as conditional sequence generation and adopt a geometry-conditioned diffusion framework that synthesizes continuous trajectories directly from raw 3D point clouds. Our method produces temporally ordered trajectory chunks and avoids post-hoc heuristic ordering or stitching in prior learning-based methods via simple sequential concatenation, improving sequence-level consistency. A single shared policy generalizes across different geometries without category-specific architectures. Extensive benchmarks show substantial gains over prior learning-based baselines: 98.2\% lower point-wise Chamfer Distance (lower is better), 97.0\% lower jerk (smoother trajectories), and +67.5 percentage points overlapping surface coverage on average.

Figures

Figures reproduced from arXiv: 2510.03011 by the authors.

Figure 1
Figure 1. Overview of the 3DCovDiffusion framework: the policy learns from [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of the 3DCovDiffusion architecture. First, input point clouds are passed through the geometry encoder, which extracts a global observation [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison of trajectory predictions for four object categories. Grey: Point Cloud, Red: ground-truth(GT), Blue: 3DCovDiffusion (Ours). [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative coverage comparison across object categories. Columns [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

36 extracted references · 5 linked inside Pith

  1. [1]

    An algorithmic perspective on imitation learning,

    T. Osa, J. Pajarinen, G. Neumann, J. A. Bagnell, P. Abbeel, J. Peters et al., “An algorithmic perspective on imitation learning,”Foundations and Trends®in Robotics, vol. 7, no. 1-2, pp. 1–179, 2018

    2018

  2. [2]

    Paintnet: Unstructured multi- path learning from 3d point clouds for robotic spray painting,

    G. Tiboni, R. Camoriano, and T. Tommasi, “Paintnet: Unstructured multi- path learning from 3d point clouds for robotic spray painting,” in2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2023, pp. 3857–3864

    2023

  3. [3]

    3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations,

    Y. Ze, G. Zhang, K. Zhang, C. Hu, M. Wang, and H. Xu, “3d diffusion policy: Generalizable visuomotor policy learning via simple 3d representations,” inProceedings of Robotics: Science and Systems (RSS), 2024

    2024

  4. [4]

    Diffusion policy: Visuomotor policy learning via ac- tion diffusion,

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y. Du, B. Burchfiel, R. Tedrake, and S. Song, “Diffusion policy: Visuomotor policy learning via ac- tion diffusion,”The International Journal of Robotics Research, p. 02783649241273668, 2023

    2023

  5. [5]

    Maskplanner: Learning- based object-centric motion generation from 3d point clouds,

    G. Tiboni, R. Camoriano, and T. Tommasi, “Maskplanner: Learning- based object-centric motion generation from 3d point clouds,”arXiv preprint arXiv:2502.18745, 2025

    arXiv 2025

  6. [6]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems, vol. 33, pp. 6840– 6851, 2020

    2020

  7. [7]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” inInternational Conference on Learning Representations, 2021. [Online]. Available: .https://openreview.net/forum?id=St1giarCHLP

    2021

  8. [8]

    Deep generative models in robotics: A survey on learning from multimodal demonstrations,

    J. Urain, A. Mandlekar, Y. Du, M. Shafiullah, D. Xu, K. Fragkiadaki, G. Chalvatzaki, and J. Peters, “Deep generative models in robotics: A survey on learning from multimodal demonstrations,”arXiv preprint arXiv:2408.04380, 2024

    arXiv 2024

Show all 36 references
  1. [9]

    Automated robot trajectory planning for spray painting of free-form surfaces in automotive manufacturing,

    H. Chen, W. Sheng, N. Xi, M. Song, and Y. Chen, “Automated robot trajectory planning for spray painting of free-form surfaces in automotive manufacturing,” inProceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No. 02CH37292), vol. 1. IEEE, 2002, pp. 450–455

    2002

  2. [10]

    Automatic trajectory generation for robotic painting application,

    X. Li, O. A. Landsnes, H. Chen, M. Sudarshan, T. A. Fuhlbrigge, and M.- A. Rege, “Automatic trajectory generation for robotic painting application,” inISR 2010 (41st International Symposium on Robotics) and ROBOTIK 2010 (6th German Conference on Robotics). VDE, 2010, pp. 1–6

    2010

  3. [11]

    Automated cad-guided robot path planning for spray painting of compound surfaces,

    W. Sheng, N. Xi, M. Song, Y. Chen, and P. MacNeille, “Automated cad-guided robot path planning for spray painting of compound surfaces,” inProceedings. 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2000) (Cat. No.00CH37113), vol. 3, 2000, pp. 1...

    2000

  4. [12]

    Generating optimized trajectories for robotic spray painting,

    D. Gleeson, S. Jakobsson, R. Salman, F. Ekstedt, N. Sandgren, F. Edelvik, J. S. Carlson, and B. Lennartson, “Generating optimized trajectories for robotic spray painting,”IEEE Transactions on Automation Science and Engineering, vol. 19, no. 3, pp. 1380–1391, 2022

    2022

  5. [13]

    Uniform coverage of automotive surface patches,

    P. Atkar, A. L. Greenfield, D. C. Conner, H. Choset, and A. Rizzi, “Uniform coverage of automotive surface patches,”International Journal of Robotics Research, vol. 24, no. 11, pp. 883 – 898, November 2005

    2005

  6. [14]

    Trajectory planning for spray painting robot based on point cloud slicing technique,

    W. Chen, X. Li, H. Ge, L. Wang, and Y. Zhang, “Trajectory planning for spray painting robot based on point cloud slicing technique,”Electronics, vol. 9, no. 6, p. 908, 2020

    2020

  7. [15]

    Incremental approach for trajectory generation of spray painting robot,

    M. V. Andulkar and S. S. Chiddarwar, “Incremental approach for trajectory generation of spray painting robot,”Industrial Robot: An International Journal, vol. 42, no. 3, pp. 228–241, 2015

    2015

  8. [16]

    Generative adversarial imitation learning,

    J. Ho and S. Ermon, “Generative adversarial imitation learning,”Advances in neural information processing systems, vol. 29, 2016

    2016

  9. [17]

    Fast multi-robot motion planning via imitation learning of mixed-integer programs,

    M. Srinivasan, A. Chakrabarty, R. Quirynen, N. Yoshikawa, T. Mariyama, and S. Di Cairano, “Fast multi-robot motion planning via imitation learning of mixed-integer programs,”IFAC-PapersOnLine, vol. 54, no. 20, pp. 598–604, 2021

    2021

  10. [18]

    A structured prediction approach for robot imitation learning,

    A. Duan, I. Batzianoulis, R. Camoriano, L. Rosasco, D. Pucci, and A. Billard, “A structured prediction approach for robot imitation learning,” The International Journal of Robotics Research, vol. 43, no. 2, pp. 113– 133, 2024

    2024

  11. [19]

    Denoise and contrast for category agnostic shape completion,

    A. Alliegro, D. Valsesia, G. Fracastoro, E. Magli, and T. Tommasi, “Denoise and contrast for category agnostic shape completion,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 4629–4638

    2021

  12. [20]

    Pointnet++ grasping: Learning an end-to-end spatial grasp generation algorithm from sparse point clouds,

    P. Ni, W. Zhang, X. Zhu, and Q. Cao, “Pointnet++ grasping: Learning an end-to-end spatial grasp generation algorithm from sparse point clouds,” in2020 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2020, pp. 3619–3625

    2020

  13. [21]

    A point set generation network for 3d object reconstruction from a single image,

    H. Fan, H. Su, and L. J. Guibas, “A point set generation network for 3d object reconstruction from a single image,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 605–613

    2017

  14. [22]

    Rt-1: Robotics transformer for real-world control at scale,

    A. Brohan, N. Brown, J. Carbajal, Y. Chebotar, J. Dabis, C. Finn, K. Gopalakrishnan, K. Hausman, A. Herzog, J. Hsuet al., “Rt-1: Robotics transformer for real-world control at scale,” inProceedings of Robotics: Science and Systems (R:SS), 2022

    2022

  15. [23]

    Human motion diffusion model,

    G. Tevet, S. Raab, B. Gordon, Y. Shafir, D. Cohen-Or, and A. H. Bermano, “Human motion diffusion model,”arXiv preprint arXiv:2209.14916, 2022

    2022 arXiv

  16. [24]

    Motion planning diffusion: Learning and adapting robot motion planning with diffusion models,

    J. Carvalho, A. T. Le, P. Kicki, D. Koert, and J. Peters, “Motion planning diffusion: Learning and adapting robot motion planning with diffusion models,”IEEE Transactions on Robotics, 2025

    2025

  17. [25]

    Edmp: Ensemble-of-costs-guided diffusion for motion planning,

    K. Saha, V. Mandadi, J. Reddy, A. Srikanth, A. Agarwal, B. Sen, A. Singh, and M. Krishna, “Edmp: Ensemble-of-costs-guided diffusion for motion planning,” in2024 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2024, pp. 10351–10358

    2024

  18. [26]

    Cold diffusion on the replay buffer: Learning to plan from known good states,

    Z. Wang, T. Oba, T. Yoneda, R. Shen, M. Walter, and B. C. Stadie, “Cold diffusion on the replay buffer: Learning to plan from known good states,” inConference on Robot Learning. PMLR, 2023, pp. 3277–3291

    2023

  19. [27]

    M 2 diffuser: Diffusion-based trajectory optimization for mobile manipulation in 3d scenes,

    S. Yan, Z. Zhang, M. Han, Z. Wang, Q. Xie, Z. Li, Z. Li, H. Liu, X. Wang, and S.-C. Zhu, “M 2 diffuser: Diffusion-based trajectory optimization for mobile manipulation in 3d scenes,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    2025

  20. [28]

    Consistency policy: Accelerated visuomotor policies via consistency distillation,

    A. Prasad, K. Lin, J. Wu, L. Zhou, and J. Bohg, “Consistency policy: Accelerated visuomotor policies via consistency distillation,”arXiv preprint arXiv:2405.07503, 2024

    2024 arXiv

  21. [29]

    Streaming diffusion policy: Fast policy synthesis with variable noise diffusion models,

    S. H. Høeg, Y. Du, and O. Egeland, “Streaming diffusion policy: Fast policy synthesis with variable noise diffusion models,”arXiv preprint arXiv:2406.04806, 2024

    2024 arXiv

  22. [30]

    Pcn: Point completion network,

    W. Yuan, T. Khot, D. Held, C. Mertz, and M. Hebert, “Pcn: Point completion network,” in2018 international conference on 3D vision (3DV). IEEE, 2018, pp. 728–737

    2018

  23. [31]

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

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 652–660

    2017

  24. [32]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space,

    C. R. Qi, L. Yi, H. Su, and L. J. Guibas, “Pointnet++: Deep hierarchical feature learning on point sets in a metric space,”Advances in neural information processing systems, vol. 30, 2017

    2017

  25. [33]

    Point transformer,

    H. Zhao, L. Jiang, J. Jia, P. H. Torr, and V. Koltun, “Point transformer,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 16259–16268

    2021

  26. [34]

    Paintrl: Coverage path planning for industrial spray painting with reinforcement learning,

    P. Yang, P. Meißner, and T. Kröger, “Paintrl: Coverage path planning for industrial spray painting with reinforcement learning,” inProceedings of the Conference Name, 2019

    2019

  27. [35]

    A point set generation network for 3d object reconstruction from a single image,

    H. Fan, H. Su, and L. Guibas, “A point set generation network for 3d object reconstruction from a single image,” in2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 2463–2471

    2017

  28. [36]

    Pointnet/pointnet++ pytorch,

    X. Yan, “Pointnet/pointnet++ pytorch,”.https://github .com/yanx27/ Pointnet_Pointnet2_pytorch, 2019

    2019

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.