Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

One human demonstration can teach a robot an object's joints, their operation order, and their states over time.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-03 05:15 UTC pith:N25S2G3O

load-bearing objection PokeNet's core articulation-estimation story is plausible and worth refereeing, but its headline novelty — manipulation order — is never actually measured. the 4 major comments →

arxiv 2602.02741 v2 pith:N25S2G3O submitted 2026-02-02 cs.RO

PokeNet: Learning Kinematic Models of Articulated Objects from Human Observations

Anmol Gupta , Weiwei Gu , Omkar Patil , Jun Ki Lee , Nakul Gopalan This is my paper
classification cs.RO
keywords articulated objectskinematic modelingjoint parameter estimationmanipulation orderhuman demonstrationpoint cloud videoset predictionmulti-DoF
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

PokeNet claims that a robot can learn how an unfamiliar articulated object moves—its joints, their types and axes, how far each joint is opened at every moment, and the order in which joints must be operated—from nothing more than a single point-cloud video of a person manipulating the object. The method needs no prior knowledge of the object's category, joint count, or joint types. If correct, this would let a robot watch one human demonstration and then manipulate a novel object itself, including multi-step objects like dishwashers whose door must open before the rack pulls out. The paper tests this in simulation and on real kitchen appliances, reporting large accuracy gains over prior single-view and reconstruction-based approaches.

Core claim

PokeNet treats articulation modeling as a set prediction problem: a transformer decoder with a fixed number of learnable queries outputs a set of joint 'slots', each carrying a confidence score, a joint type (revolute or prismatic), an axis direction, an anchor point on the axis, and a manipulation-order score. A permutation-invariant matching loss aligns predicted slots to ground-truth joints, so the model can output a variable number of joints without knowing the count in advance. An auxiliary decoder predicts, for every retained slot and every observed frame, the joint's displacement—angular for revolute joints, linear for prismatic joints—so the model recovers the full time-varying state

What carries the argument

The central mechanism is a slot-based set predictor with permutation-invariant matching. A point-cloud encoder followed by two transformer encoders compresses each frame spatially and the whole sequence temporally; then a decoder with a fixed set of learnable queries specializes each query into a joint slot containing confidence, joint type, normalized axis direction, anchor point, and order score. Hungarian matching aligns predicted slots to ground-truth joints, removing any assumption about the number of joints. An auxiliary decoder outputs per-slot, per-frame joint states as (sin θ, cos θ) for revolute joints and linear displacement for prismatic joints, providing a wraparound-free angula

Load-bearing premise

The demonstration must actually move every joint the model is expected to recover, and the single camera view must reveal each joint's motion at some point in the sequence; joints that are never actuated or remain occluded throughout cannot be estimated.

What would settle it

Record a demonstration of a multi-joint object in which one joint is never moved and another stays occluded for the entire video, then run PokeNet. If the model assigns high confidence to a slot for a joint that was never revealed, or fails to lower confidence for a joint that never moved, the claim that occluded or un-actuated joints are handled collapses. Conversely, if it correctly omits those joints, actuation and visibility are confirmed as necessary. A second test: demonstrate the same object twice with different joint operation orders and check whether the predicted order scores track t

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • A robot can acquire a usable articulation model of an unseen object from a single demonstration, with no prior object category, joint-count, or joint-type knowledge.
  • Multi-step objects like dishwashers can be manipulated in the correct sequence because the predicted order score encodes the demonstrated operation order.
  • Joints that are occluded or fully shut when static can still be modeled as long as the human demonstration reveals them during the interaction.
  • The learned articulation model can be plugged into a motion planner: the paper demonstrates a robot opening a refrigerator using PokeNet's inferred parameters.
  • The released real-world dataset of 5,500 annotated human–object interaction sequences provides a resource for training and evaluating articulation models beyond this paper.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Because PokeNet learns from natural human manipulation rather than scripted robot actions, it may be possible to train it on unlabeled in-the-wild videos (household footage, online clips) using self-supervision, though the current training data is fully annotated.
  • The inferred manipulation order is effectively a weak task plan: the ordered joint sequence describes how to reach the demonstrated configuration. Extending this to contact-point detection and collision-aware planning is a natural next step the paper leaves open.
  • The slot-based formulation suggests a direct extension to additional joint types (e.g., helical or multi-axis joints) by adding new type classes and corresponding state encodings; the paper only handles revolute and prismatic joints.
  • PokeNet's time-aligned per-joint state predictions could serve as supervision for imitation learning of manipulation skills, since they already provide the full joint trajectory from a single demonstration.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes PokeNet, an end-to-end transformer-based framework that takes a sequence of single-view point clouds of a human manipulating an unknown object and predicts a set of joint slots. Each slot contains a confidence score, joint type (revolute/prismatic), axis direction, anchor point, and an order score; an auxiliary decoder predicts per-frame joint states. Training uses a DETR-style set prediction loss with Hungarian matching, plus losses for confidence, type, axis, anchor, order (L1 and ranking), and joint state. The method is evaluated on a simulated PartNet-Mobility dataset with 11 training and 4 held-out categories, and on a newly collected real-world dataset of 5,500 sequences for microwave, fridge, drawer, and dishwasher, plus two unseen real objects. The reported results show lower joint axis orientation error, axis displacement error, and joint state error than ScrewNet and GAPartNet, and a single qualitative real-robot demo of opening a refrigerator.

Significance. If the results hold, the contribution is meaningful: recovering articulation models from a single human demonstration without category priors, including occluded joints and multi-DoF objects, is a step beyond single-frame methods. The released real-world dataset is a useful resource. The paper also correctly frames the set-prediction formulation as a way to avoid assuming joint count. However, the manuscript's headline novelty—manipulation-order inference—is never quantitatively evaluated, and the baseline comparisons use different input modalities without controls. These gaps are load-bearing for the central claims, so the paper requires substantial revision before the significance can be assessed.

major comments (4)
  1. [Section VI, Tables II–III; Section V-B, Eqs. (1)–(3)] Manipulation order is advertised as a core contribution (Table I, abstract, Section I), and the order score o_k is trained via Eqs. (1)–(3), yet no experiment reports order prediction accuracy. Tables II and III report only axis orientation, axis displacement, and joint state error. The sole evidence for order inference is the qualitative refrigerator demo in Fig. 3. This is load-bearing: if PokeNet cannot reliably infer the correct manipulation sequence, the claimed 'multi-DoF operation order' contribution collapses. Please add quantitative order metrics (e.g., exact-match rate, Kendall tau, or normalized order error) on the simulated and real test sets, for objects with at least two joints, and compare with baselines. Also discuss whether the model can exploit a temporal shortcut: since demonstrations move joints in a fixed order, o_k might simply encode the first time each joint moves
  2. [Section VI-A, VI-B and Tables II, III footnotes] The comparison against GAPartNet is confounded by input modality: GAPartNet is given a single partial point cloud in a partially opened state, while PokeNet receives the full point-cloud sequence of a human demonstration. The footnotes in Tables II and III state 'All objects were in partially opened state for GAPartNet,' but this setup is not described in the main text, and no PokeNet variant uses a single frame. This makes it difficult to attribute the reported improvements to the method rather than to the extra information in the demonstration sequence. Please provide a single-frame PokeNet baseline or run the baselines on the same sequence input, or otherwise explicitly ablate the effect of temporal input.
  3. [Section V-A, V-B] There are no ablations isolating the contributions of the major architectural and loss components: the temporal transformer, the DETR-style set decoder, the auxiliary state decoder, the order losses (L1 and ranking), and the confidence threshold. Given the composite loss in Eq. (4) and the many design choices, the paper should include at least a small ablation study (e.g., removing the temporal encoder, removing the rank loss, or using a fixed number of joints) to show which components are necessary for the reported performance.
  4. [Section VI-B, Fig. 3] The real-robot validation is a single qualitative example (Sawyer opening a refrigerator) with no quantitative success metrics: no number of trials, no success rate, no trajectory error, and no comparison on multiple objects. The conclusion claims 'successful manipulation on both seen and unseen object classes,' but the evidence shown is one fridge demo. Please report systematic real-robot experiments with several objects (including at least one unseen category) and quantitative metrics such as success rate, completion time, or joint-state tracking error during execution.
minor comments (6)
  1. [Section IV-A, Section VI-A] Dataset size inconsistency: Section IV-A says 110,000 simulated sequences, but Section VI-A reports 88,000 training sequences and 2,000 test sequences per category across the 11+4 categories (30,000 test sequences), totaling 118,000. Please clarify whether the '110,000' figure refers only to training data or whether the test splits overlap differently.
  2. [Abstract, Section I, Section VIII] Reported improvement is inconsistent: the abstract states 'over 27%,' the introduction says 'up to 25%... and by 30%,' and the conclusion says 'by25%... and by30%.' Please unify these numbers and specify the exact averaging across categories and metrics.
  3. [Tables II, III] The footnote 'All objects were in partially opened state for GAPartNet' is important. It should be described in the experimental setup section, explaining why this choice was made and whether it favors or disfavors the baselines.
  4. [Section VI-A] The extension of ScrewNet to multi-joint objects is mentioned but not described. Since the original ScrewNet is designed for single-DoF objects, please explain the extension and any changes to its training/evaluation.
  5. [Section III] The notation for joint state y_{t,k} uses (sin θ, cos θ, 0) for revolute and (0,0,ρ) for prismatic. The text says this is 'wraparound-free,' but no table reports state error separately for revolute and prismatic. It would help to report these separately, since combining degrees and centimeters is not directly comparable.
  6. [Section VII] The limitation statement correctly notes that PokeNet does not estimate contact points and assumes obstacle-free motion. This is fine, but the conclusion's phrase 'successful manipulation' should be tempered given these limitations and the limited real-robot evaluation.

Circularity Check

0 steps flagged

No significant circularity: supervised learning with held-out test data; the unquantified manipulation-order claim is an evaluation gap, not circularity.

full rationale

PokeNet is a supervised learning system whose predictions are trained against ground-truth labels and evaluated on held-out test splits (Tables II and III). No parameter is fitted to the test data, and the joint-axis, displacement, and state errors are reported on unseen instances and held-out categories, so the central quantitative claims are not equivalent to the training inputs by construction. The manipulation-order score is supervised with an L1 loss and a pairwise ranking loss (Eqs. 2–3) against ground-truth demonstration order; Eq. (1) is only a label-normalization map, not an injection of the model output into the target. Sorting retained slots by the predicted order score at inference is a standard decoding choice, not a circular derivation. The references to DETR, PointNet++, and transformer architectures are external, standard, and non-load-bearing; there is no self-citation chain or imported uniqueness theorem used to force the method. The limitations stated in Section VII (no contact points, no obstacle avoidance, no full geometry) are acknowledged weaknesses rather than hidden circular assumptions. The skeptical concern that manipulation-order accuracy is never quantitatively measured is a legitimate completeness/validity weakness, but absence of an evaluation does not make the derivation circular. Overall, the paper's claims are testable, externally benchmarked, and self-contained in the sense required for a non-circularity finding.

Axiom & Free-Parameter Ledger

4 free parameters · 5 axioms · 0 invented entities

The model's success depends on the data-generation assumptions (rigid links, bounded joint count, a demonstration that moves and reveals all joints) and on the accuracy of the ground-truth labels. Hyperparameters governing the loss and inference are not reported, which is a reproducibility concern.

free parameters (4)
  • Maximum joint slots K
    The set decoder uses K learnable slots; the value is a hand-chosen upper bound on the number of joints, not specified in the paper.
  • Confidence threshold mu
    Slots with c_k > mu are retained; mu is a hyperparameter chosen by hand and not reported.
  • Ranking margin m
    Eq. (3) uses margin m in the pairwise order hinge loss; not reported.
  • Loss weights lambda_*
    Eq. (4) combines seven losses with weights; no values are given.
axioms (5)
  • domain assumption Object links are rigid and the number of joints is at most K.
    Stated in Section III Problem Formulation; the model cannot handle non-rigid parts or more than K joints.
  • domain assumption A single-view point cloud sequence of a human demonstration contains sufficient visual information to recover all joint axes, states, and order.
    Central to the method; if occlusions persist or the view is uninformative, the model fails. The paper assumes the demonstration reveals occluded joints during interaction.
  • domain assumption The demonstration actuates every joint of interest.
    Section IV-B says objects were manipulated by opening/closing movable parts; if a joint is not moved, it cannot be inferred.
  • domain assumption Ground-truth parameters from the simulator and ArUco tracking are accurate.
    The supervised training and evaluation rely on these annotations; marker tracking errors would propagate.
  • standard math Hungarian matching gives the minimum-cost alignment between slots and ground-truth joints.
    The DETR-style loss (Section V-B) depends on this known algorithm's correctness.

pith-pipeline@v1.3.0-alltime-deepseek · 10695 in / 14721 out tokens · 140458 ms · 2026-08-03T05:15:51.300541+00:00 · methodology

0 comments
read the original abstract

Articulation modeling enables robots to learn joint parameters of articulated objects for effective manipulation which can then be used downstream for skill learning or planning. Existing approaches often rely on prior knowledge about the objects, such as the number or type of joints. Some of these approaches also fail to recover occluded joints that are only revealed during interaction. Others require large numbers of multi-view images for every object, which is impractical in real-world settings. Furthermore, prior works neglect the order of manipulations, which is essential for many multi-DoF objects where one joint must be operated before another, such as a dishwasher. We introduce PokeNet, an end-to-end framework that estimates articulation models from a single human demonstration without prior object knowledge. Given a sequence of point cloud observations of a human manipulating an unknown object, PokeNet predicts joint parameters, infers manipulation order, and tracks joint states over time. PokeNet outperforms existing state-of-the-art methods, improving joint axis and state estimation accuracy by an average of over 27% across diverse objects, including novel and unseen categories. We demonstrate these gains in both simulation and real-world environments.

Figures

Figures reproduced from arXiv: 2602.02741 by Anmol Gupta, Jun Ki Lee, Nakul Gopalan, Omkar Patil, Weiwei Gu.

Figure 1
Figure 1. Figure 1: We propose a novel framework that learns the joint parameters and manipulation order of articulated objects directly [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of PokeNet. Our model takes a sequence of point clouds as input. Each frame is encoded with PointNet++ to extract spatial features, and [CLS] tokens are passed through a transformer encoder to capture temporal dependencies. A DETR style joint decoder with learnable queries attends to these temporal features to predict an ordered kinematic slot model, including joint type, axis direction, anchor po… view at source ↗
Figure 3
Figure 3. Figure 3: This figure shows Sawyer robot manipulating the two joints of the fridge in the order of demonstration estimated by [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Comparison of PokeNet and GAPartNet on differ [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. QDTraj: Exploration of Diverse Trajectory Primitives for Articulated Objects Robotic Manipulation

    cs.RO 2026-04 unverdicted novelty 6.0

    QDTraj uses Quality-Diversity algorithms with sparse rewards to produce at least five times more diverse high-performing trajectories for articulated object manipulation than compared methods, validated across 30 obje...

Reference graph

Works this paper leans on

29 extracted references · 12 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Learning to generalize kinematic models to novel objects,

    B. Abbatematteo, S. Tellex, and G. D. Konidaris, “Learning to generalize kinematic models to novel objects,” inConference on Robot Learning, 2019. [Online]. Available: https://api.semanticscholar.org/ CorpusID:204843155

    2019

  2. [2]

    Flowbot++: Learning generalized articulated objects manipulation via articulation projection,

    H. Zhang, B. Eisner, and D. Held, “Flowbot++: Learning generalized articulated objects manipulation via articulation projection,” 2024

    2024

  3. [3]

    Flowbot3d: Learning 3d artic- ulation flow to manipulate articulated objects,

    B. Eisner*, H. Zhang*, and D. Held, “Flowbot3d: Learning 3d artic- ulation flow to manipulate articulated objects,” inRobotics: Science and Systems (RSS), 2022

    2022

  4. [4]

    Urdformer: A pipeline for constructing articulated simulation environments from real-world images,

    Z. Chen, A. Walsman, M. Memmel, K. Mo, A. Fang, K. Vemuri, A. Wu, D. Fox, and A. Gupta, “Urdformer: A pipeline for constructing articulated simulation environments from real-world images,”arXiv preprint arXiv:2405.11656, 2024

    Pith/arXiv arXiv 2024

  5. [5]

    Gapartnet: Cross-category domain-generalizable object perception and manipulation via generalizable and actionable parts,

    H. Geng, H. Xu, C. Zhao, C. Xu, L. Yi, S. Huang, and H. Wang, “Gapartnet: Cross-category domain-generalizable object perception and manipulation via generalizable and actionable parts,”arXiv preprint arXiv:2211.05272, 2022

    Pith/arXiv arXiv 2022

  6. [6]

    Screwnet: Category- independent articulation model estimation from depth images using screw theory,

    A. Jain, R. Lioutikov, C. Chuck, and S. Niekum, “Screwnet: Category- independent articulation model estimation from depth images using screw theory,” in2021 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2021, pp. 13 670–13 677

    2021

  7. [7]

    Distributional depth-based estimation of object articulation models,

    A. Jain, S. Giguere, R. Lioutikov, and S. Niekum, “Distributional depth-based estimation of object articulation models,” inConference on Robot Learning. PMLR, 2022, pp. 1611–1621

    2022

  8. [8]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “Nerf: Representing scenes as neural radiance fields for view synthesis,” 2020. [Online]. Available: https://arxiv.org/abs/2003.08934

    Pith/arXiv arXiv 2020

  9. [9]

    Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,

    P. Wang, L. Liu, Y . Liu, C. Theobalt, T. Komura, and W. Wang, “Neus: Learning neural implicit surfaces by volume rendering for multi-view reconstruction,”arXiv preprint arXiv:2106.10689, 2021

    Pith/arXiv arXiv 2021

  10. [10]

    3d gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimk ¨uhler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,”ACM Transactions on Graphics, vol. 42, no. 4, July 2023. [Online]. Available: https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/

    2023

  11. [11]

    Screwsplat: An end-to-end method for articulated object recognition,

    S. Kim, J. Ha, Y . H. Kim, Y . Lee, and F. C. Park, “Screwsplat: An end-to-end method for articulated object recognition,”arXiv preprint arXiv:2508.02146, 2025

    Pith/arXiv arXiv 2025

  12. [12]

    Artgs:3d gaussian splatting for interactive visual-physical modeling and manipulation of articulated objects,

    Q. Yu, X. Yuan, Y . jiang, J. Chen, D. Zheng, C. Hao, Y . You, Y . Chen, Y . Mu, L. Liu, and C. Lu, “Artgs:3d gaussian splatting for interactive visual-physical modeling and manipulation of articulated objects,” 2025. [Online]. Available: https://arxiv.org/abs/2507.02600

    Pith/arXiv arXiv 2025

  13. [13]

    Artgs: Building interactable replicas of complex articulated objects via gaussian splatting,

    Y . Liu, B. Jia, R. Lu, J. Ni, S.-C. Zhu, and S. Huang, “Artgs: Building interactable replicas of complex articulated objects via gaussian splatting,” 2025. [Online]. Available: https://arxiv.org/abs/2502.19459

    Pith/arXiv arXiv 2025

  14. [14]

    Robot see robot do: Imitating articulated object manipulation with monocular 4d reconstruction,

    J. Kerr, C. M. Kim, M. Wu, B. Yi, Q. Wang, K. Goldberg, and A. Kanazawa, “Robot see robot do: Imitating articulated object manipulation with monocular 4d reconstruction,” 2024. [Online]. Available: https://arxiv.org/abs/2409.18121

    Pith/arXiv arXiv 2024

  15. [15]

    Category-level articulated object pose estimation,

    X. Li, H. Wang, L. Yi, L. Guibas, A. L. Abbott, and S. Song, “Category-level articulated object pose estimation,” 2020. [Online]. Available: https://arxiv.org/abs/1912.11913

    Pith/arXiv arXiv 2020

  16. [16]

    Deep part induction from articulated object pairs,

    L. Yi, H. Huang, D. Liu, E. Kalogerakis, H. Su, and L. Guibas, “Deep part induction from articulated object pairs,”ACM Trans. Graph., vol. 37, no. 6, dec 2018. [Online]. Available: https: //doi.org/10.1145/3272127.3275027

    arXiv 2018

  17. [17]

    Manipulating articulated objects with interac- tive perception,

    D. Katz and O. Brock, “Manipulating articulated objects with interac- tive perception,” in2008 IEEE International Conference on Robotics and Automation, 2008, pp. 272–277

    2008

  18. [18]

    Interactive segmentation, tracking, and kinematic modeling of unknown 3d artic- ulated objects,

    D. Katz, M. Kazemi, J. A. D. Bagnell, and A. T. Stentz, “Interactive segmentation, tracking, and kinematic modeling of unknown 3d artic- ulated objects,” inProceedings of (ICRA) International Conference on Robotics and Automation, May 2013, pp. 5003 – 5010

    2013

  19. [19]

    An integrated approach to visual perception of articulated objects,

    R. Mart ´ın-Mart´ın, S. H ¨ofer, and O. Brock, “An integrated approach to visual perception of articulated objects,” in2016 IEEE International Conference on Robotics and Automation (ICRA), 2016, pp. 5091– 5097

    2016

  20. [20]

    Structure from action: Learning interactions for articulated object 3d structure discovery,

    N. Nie, S. Y . Gadre, K. Ehsani, and S. Song, “Structure from action: Learning interactions for articulated object 3d structure discovery,” arxiv, 2022

    2022

  21. [21]

    PARIS: Part-level recon- struction and motion analysis for articulated objects,

    J. Liu, A. Mahdavi-Amiri, and M. Savva, “PARIS: Part-level recon- struction and motion analysis for articulated objects,” inProceedings of the IEEE International Conference on Computer Vision (ICCV), 2023

    2023

  22. [22]

    Neural implicit representation for building digital twins of unknown articulated objects,

    Y . Weng, B. Wen, J. Tremblay, V . Blukis, D. Fox, L. Guibas, and S. Birchfield, “Neural implicit representation for building digital twins of unknown articulated objects,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 3141–3150

    2024

  23. [23]

    SAPIEN: A simulated part-based interactive environment,

    F. Xiang, Y . Qin, K. Mo, Y . Xia, H. Zhu, F. Liu, M. Liu, H. Jiang, Y . Yuan, H. Wang, L. Yi, A. X. Chang, L. J. Guibas, and H. Su, “SAPIEN: A simulated part-based interactive environment,” inThe IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2020

    2020

  24. [24]

    Shapenet: An information-rich 3d model repository,

    A. X. Chang, T. Funkhouser, L. Guibas, P. Hanrahan, Q. Huang, Z. Li, S. Savarese, M. Savva, S. Song, H. Suet al., “Shapenet: An information-rich 3d model repository,”arXiv preprint arXiv:1512.03012, 2015

    Pith/arXiv arXiv 2015

  25. [25]

    PartNet: A large-scale benchmark for fine-grained and hierarchical part-level 3D object understanding,

    K. Mo, S. Zhu, A. X. Chang, L. Yi, S. Tripathi, L. J. Guibas, and H. Su, “PartNet: A large-scale benchmark for fine-grained and hierarchical part-level 3D object understanding,” inThe IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2019

    2019

  26. [26]

    Automatic generation and detection of highly reliable fiducial markers under occlusion,

    S. Garrido-Jurado, R. Mu ˜noz-Salinas, F. J. Madrid-Cuevas, and M. J. Mar´ın-Jim´enez, “Automatic generation and detection of highly reliable fiducial markers under occlusion,”Pattern Recognition, vol. 47, no. 6, pp. 2280–2292, 2014

    2014

  27. [27]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” 2020. [Online]. Available: https://arxiv.org/abs/2005.12872

    Pith/arXiv arXiv 2020

  28. [28]

    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,” 2017. [Online]. Available: https://arxiv.org/abs/1706.02413

    Pith/arXiv arXiv 2017

  29. [29]

    Attention is all you need,

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

    2017