Pith. sign in

REVIEW 5 major objections 6 minor 29 references

Learning to Push, Group, and Grasp: A Diffusion Policy Approach for Multi-Object Delivery

T0 review · 5 major / 6 minor · reviewed 2026-08-08 · deepseek-v4-flash

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

Pith's one-line read A diffusion policy trained via teleoperated imitation learning learns to push, group, and simultaneously grasp multiple objects, matching a rule-based planner in completion rate while being faster.

desk verdict Solid empirical application of diffusion policy to push-group-grasp; a real but fixable gap in the gripper-state mapping warrants revision before acceptance. read the letter →

arxiv 2502.08452 v3 pith:CG5DICZB submitted 2025-02-12 cs.RO

Takahiro Yonemaru , Weiwei Wan , Tatsuki Nishimura , Kensuke Harada This is my paper · ORCID
classification cs.RO
keywords multi-objectgraspingdiffusionpolicyimitationlearningteleoperationpushinggroupingrobotmanipulationclosed-loopcontrol
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 claims that a diffusion policy trained on teleoperated human demonstrations can learn the full push-group-grasp routine for multi-object delivery without hand-crafted rules or reward design. This matters because rule-based planners need precomputed global plans that break when objects shift, while the learned policy reacts in closed loop and executes faster, achieving 1.77 objects per grasp versus 1.44 for the rule-based baseline. The work also shows the routine transfers to new object shapes with modest fine-tuning, and that performance scales with demonstration data.

What carries the argument

The central object is the diffusion policy, a visuomotor policy that learns to generate multi-step action sequences by iteratively denoising a noisy action sequence conditioned on observations. A U-Net core predicts the noise, ResNet-18 encoders process global and local camera frames, and the robot state (pose plus binary gripper state) is appended to the visual features. The policy is trained by minimizing the mean squared error between added noise and predicted noise, and at inference it executes the first 8 predicted actions in closed loop before re-predicting. The training data comes from a vision-based teleoperation system that maps WiLoR hand skeletons to the robot's end-effector pose and gripper state.

What would settle it

Take the 300-demonstration policy and run it on a 6-cuboid scene where every object is initially placed beyond the gripper's maximum opening width, so no direct grasp is possible. The paper's claim says the robot will still group and grasp with a completion rate near the trained 70%; if the rate falls below 30%, the policy is chaining memorized motions rather than generalizing the push-group-grasp skill.

Watch

Extended reading notes

Core claim

The central claim is that a diffusion policy, trained by imitation on human teleoperation demonstrations, can adaptively generate sequences of pushing, grouping, and simultaneous multi-object grasping actions. The robot takes two past observations (global and local camera frames plus robot state) and predicts an 8-step action chunk by denoising a Gaussian-corrupted action sequence with a U-Net. In experiments with 3D-printed cuboids, the policy trained on 300 demonstrations reached 85% completion for 4 objects and 70% for 6 objects, with a delivery rate of 86% and 1.77 objects per grasp, comparing favorably to the rule-based analytical method's 70% completion and 1.44 objects per grasp. The paper further argues that a small amount of fine-tuning adapts the policy to snack bars and unseen shapes, and that failures concentrate in pushing and stagnation.

Load-bearing premise

The demonstrations encode good strategies only if the hand-skeleton teleoperation maps human hand motion to the robot accurately; the gripper state is binarized with hand-chosen thresholds (0.8 and 0.1) and the end-effector orientation is fixed vertically downward, so noisy tracking or a lost degree of dexterity would be inherited by the policy.

Editorial extensions

If this is right

  • With 300 demonstrations the policy reaches 1.77 objects per grasp on 6-object scenes, beating the rule-based analytical method's 1.44, so a single learned policy can outperform handcrafted routines on grasping efficiency.
  • Increasing demonstration count from 100 to 300 raises completion rate from 50% to 85% on 4-object tasks and from 10% to 70% on 6-object tasks, making data volume a direct lever on performance.
  • Because the policy predicts action chunks and re-plans every 8 steps in closed loop, it can adapt when pushed objects shift, whereas the rule-based method's precomputed global plans become invalid after displacements.
  • Fine-tuning with 50 additional demonstrations transfers the routine to new objects, yielding 25% completion on snack bars and 40-60% completion on L-shapes, cubes, and hexagons, suggesting the learned routine is not object-specific.

Reading between the lines

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

  • If this holds, multi-object grasping no longer needs a geometric planner, and any task that decomposes into push-group-grasp could be learned from a few hundred demonstrations, lowering the entry cost for warehouse picking.
  • The fixed downward orientation and binary gripper suggest the policy operates in a reduced action space; extending the teleoperation mapping to full wrist rotations and continuous gripper force might let the same diffusion framework handle tilted or non-planar objects.
  • Because failures concentrate in pushing and stagnation, a hybrid that uses the diffusion policy to choose goals and an analytical push controller to execute them could combine the adaptability of learning with the reliability of geometric planning.
  • A direct extension would be to use the same teleoperation pipeline to collect demonstrations for deformable or granular objects, where rule-based grouping is even harder, and test whether the learned grouping generalizes.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. This paper proposes an imitation-learning pipeline for multi-object delivery: a human teleoperates a UR3/Robotiq gripper via WiLoR hand-skeleton tracking; the collected demonstrations (up to 300) train a diffusion policy that outputs gripper poses and binary open/close actions; the policy is evaluated on 3D-printed cuboids under varying data sizes and object counts, against a rule-based planner, after fine-tuning on snack bars, and on unseen shapes. The central claim is that the learned policy can generate pushing, grouping, and simultaneous multi-object grasping strategies adaptively, and that more training data improves performance.

Significance. If the central claim holds, the paper's main value is a demonstration that an off-the-shelf Diffusion Policy can acquire the push-group-grasp routine from teleoperated demonstrations without reward engineering, and that it can beat a rule-based planner in wall-clock efficiency while matching completion rate (70%, 1.77 vs 1.44 objects/grasp, 71 s vs 149 s for 6 objects). The experimental coverage is a strength: four experiment groups, real consumer goods, fine-tuning transfer, and a documented failure taxonomy. However, the contribution is currently supported only by point estimates from 20 (and in places 5) trials, and by a teleoperation action mapping that is incompletely specified; these gaps are load-bearing for the empirical claims. No code, data, or trained models are released, which further limits reproducibility.

major comments (5)
  1. [Section III, gripping state definition] The mapping g=0 if d>=mu_up and g=1 if d<mu_bot, with mu_up=0.8 and mu_bot=0.1, leaves every normalized thumb-index distance in [0.1,0.8) without an assigned gripper state. Because g is the only action channel that distinguishes pushing motions from grasping motions, the training labels for all demonstration frames in this interval are either silently defined by an unspecified rule or censored; either way the paper cannot be reproduced from the text alone. Please specify the behavior in the deadband, report the fraction of demonstration frames that fall in it, or release code/data that disambiguates the mapping.
  2. [Section IV, action representation and inference] The diffusion policy outputs a continuous 8xTp action sequence, but the gripper state g is binary. The paper does not state how the continuous denoised output for g is mapped to open/close commands at execution time, nor whether the training loss treats g as continuous. This is load-bearing because g determines whether the robot pushes or grasps. Please specify the discretization or threshold used at inference and describe how it interacts with the undefined deadband in Section III.
  3. [Section III, rotation mapping for the end effector] The displayed rotation matrix for Sigma_wd R_ef has determinant -1 for unit d (columns (dx,dy,0), (dy,-dx,0), (0,0,1)), so it is a reflection rather than a proper rotation. If this matrix is used literally to set the end-effector orientation, the action space is not a rigid-body pose. Please correct the sign convention, explain the coordinate convention that makes the matrix a valid rotation, and confirm that the quaternion used in s_t and a_t is derived from a proper rotation.
  4. [Section V, all experiments] Every quantitative conclusion--data-volume effect, comparison with the rule-based baseline, fine-tuning transfer, and unseen-shape generalization--rests on 20-trial or 5-trial point estimates with no confidence intervals, standard deviations, or statistical tests. For example, Table I completion rates of 85% versus 70% across data sizes and Table III's 70% baseline cannot be distinguished from sampling noise at n=20; Table V uses only 5 trials per shape. Please report per-trial distributions, confidence intervals, and appropriate significance tests, or justify why they are unnecessary for the specific claims.
  5. [Section V-B, baseline comparison] The baseline is described only as 'we improved the approach proposed by Sakamoto et al. [1] to support combined pushing and grasping for three objects,' with no algorithmic details, parameters, objective function, or failure criteria. Since the baseline is the authors' own prior work and is modified for this comparison, the fairness and reproducibility of the time-efficiency and objects-per-grasp comparison cannot be assessed. Please specify the baseline algorithm and its modifications in sufficient detail, and ideally release the planner.
minor comments (6)
  1. [Section III, text near gripping state] The sentence 'The threshold values mu_up and mu_up are chosen' should read 'mu_up and mu_bot.'
  2. [Section III, last sentence] The sentence 'We clarified the above discussion at the beginning of the experimental section' appears to be an editorial artifact and should be removed or rewritten as a proper cross-reference.
  3. [Section IV, network architecture] The hidden-layer list '[512, 1024, 2024]' is likely a typo for '[512, 1024, 2048]'; please correct it.
  4. [Table II, note] The note defines t1 for one object and then defines 't3' twice, for two and three objects; the two-object entry should be labeled t2.
  5. [Table V, note and header] The note refers to '#S' and '#F' but the table uses '#G'; also, unlike other tables, Table V has no summary row, making it hard to compare with Table VI.
  6. [References and main text, [22]] The main text cites 'Kishore et al. [22]' for the busboy problem, but reference [22] is authored by Srinivas et al.; the citation should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical evaluation uses held-out trials and external success metrics, so the central claim is self-contained.

full rationale

The paper's derivation chain is empirical and self-contained. Demonstrations are collected via visual teleoperation (Sec. III), mapped to robot positions, orientations, and a binarized gripper state; a diffusion policy is trained with the standard noise-prediction loss (Eq. 3); and the trained policy is executed closed-loop on held-out object configurations (Sec. V). The reported metrics—completion rate, delivery rate, and objects per grasp—are external counts of task outcome, not functions of the training labels or fitted network parameters, so the central claim does not reduce to its inputs by construction. The only self-referential element is the rule-based baseline [1] by overlapping authors (Sakamoto, Wan, Nishi, Harada), which the present authors extend to three-object grasps in Sec. V-B; that baseline is a comparison point, not the justification for the proposed method, and the absolute performance of the learned policy in Table I stands independently of it. A genuine gap exists in Sec. III: the binarized gripper mapping leaves normalized distances in [0.1, 0.8) undefined, so the training label for intermediate hand configurations is unspecified, and code/data are not released. This is a reproducibility and correctness concern about demonstration quality, not a circular derivation, because it does not equate a predicted quantity with an already-known input. No load-bearing self-citation, imported uniqueness theorem, or renamed empirical pattern was found.

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

The central claim is empirical and rests on data quality and representation choices. No new physical entities are introduced.

free parameters (3)
  • Gripper state binarization thresholds = µup=0.8, µbot=0.1
    Hand-chosen thresholds in Section III that convert the normalized thumb-index distance into a binary open/close state. This coarse action representation may limit the dexterity of the learned policy.
  • Maximum objects per group (operator heuristic) = 3
    Note 1 states the human operator limited grouping to three objects to avoid overloading the gripper. This heuristic is embedded in the demonstrations and thus in the learned policy, constraining its behavior.
  • Observation and action horizons = To=2, Tp=16, Ta=8
    Chosen architecture parameters in Section V that determine the context and planning horizon; To=2 means the policy sees only the current and one previous frame, which may be too short for dynamic scenes.
assumptions (3)
  • domain assumption The WiLoR-based teleoperation mapping provides accurate and low-latency hand-to-robot action conversion.
    Section III defines the mapping; if tracking is noisy or delayed, the demonstrations are poor and the policy fails.
  • domain assumption Human expert demonstrations encode effective grouping and grasping heuristics.
    The policy learns from demonstrations; suboptimal or inconsistent human strategies would directly degrade the learned policy. The operator heuristics in Note 1 are assumed to be good.
  • domain assumption Two RGB frames (To=2) from the global and local cameras are sufficient state for the policy to infer object configurations and plan multi-step actions.
    The observation space is defined in Section III and the horizon in Section V; if two frames cannot capture object motion or the state, the policy cannot plan effectively.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning to Push, Group, and Grasp: A Diffusion Policy Approach for Multi-Object Delivery." pith.science (2026). https://pith.science/paper/CG5DICZB

@misc{pith2026250208452,
  author       = {Pith},
  title        = {Pith review of: Learning to Push, Group, and Grasp: A Diffusion Policy Approach for Multi-Object Delivery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CG5DICZB}},
  note         = {Machine review of arXiv:2502.08452}
}
read the original abstract

Simultaneously grasping and delivering multiple objects can significantly enhance robotic work efficiency and has been a key research focus for decades. The primary challenge lies in determining how to push objects, group them, and execute simultaneous grasping for respective groups while considering object distribution and the hardware constraints of the robot. Traditional rule-based methods struggle to flexibly adapt to diverse scenarios. To address this challenge, this paper proposes an imitation learning-based approach. We collect a series of expert demonstrations through teleoperation and train a diffusion policy network, enabling the robot to dynamically generate action sequences for pushing, grouping, and grasping, thereby facilitating efficient multi-object grasping and delivery. We conducted experiments to evaluate the method under different training dataset sizes, varying object quantities, and real-world object scenarios. The results demonstrate that the proposed approach can effectively and adaptively generate multi-object grouping and grasping strategies. With the support of more training data, imitation learning is expected to be an effective approach for solving the multi-object grasping problem.

Figures

Figures reproduced from arXiv: 2502.08452 by the authors.

Figure 1
Figure 1. (a) Collecting human expert data using teleoperation. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Mapping relationship between hand poses detected by [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The diffusion policy takes as input the observations [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Internal mechanism of the diffusion policy network. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: An exemplary result. (a∼c) First grouping and grasp￾ing. (d) Second grasping, (e, f) Third grouping and grasping. motions, please refer to the supplementary video. B. Experiment 2: Comparison with analytical methods Second, we conducted a comparative study against a ru…
Figure 6
Figure 6. Figure 6: Snack bar trial with a fine-tuned model. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 28 canonical work pages

  1. [1]

    Efficient picking by considering simultaneous two-object grasp- ing,

    T. Sakamoto, W. Wan, T. Nishi, and K. Harada, “Efficient picking by considering simultaneous two-object grasp- ing,” in IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS) , 2021, pp. 8295–8300

    work page 2021

  2. [2]

    Multi-object grasping in the plane,

    W. C. Agboh, J. Ichnowski, K. Goldberg, and M. R. Dogar, “Multi-object grasping in the plane,” in Int. Symp. Robot. Res. (ISRR) . Springer, 2022, pp. 222–238

    work page 2022

  3. [3]

    Diffusion policy: Visuo- motor policy learning via action diffusion,

    C. Chi, Z. Xu, S. Feng, E. Cousineau, Y . Du, B. Burch- fiel, R. Tedrake, and S. Song, “Diffusion policy: Visuo- motor policy learning via action diffusion,” Int. J. Robot. Res., p. 02783649241273668, 2023

    work page 2023

  4. [4]

    Manipulation of multiple objects by two manipulators,

    Y . Aiyama, M. Minami, and T. Arai, “Manipulation of multiple objects by two manipulators,” in IEEE Int. Conf. Robot. Autom. (ICRA) , 1998, pp. 2904–2909

    work page 1998

  5. [5]

    Neighborhood equilibrium grasp for multiple objects,

    K. Harada and M. Kaneko, “Neighborhood equilibrium grasp for multiple objects,” in IEEE Int. Conf. Robot. Autom. (ICRA) , 2000, pp. 2159–2164

    work page 2000

  6. [6]

    Enveloping grasp for multiple objects,

    ——, “Enveloping grasp for multiple objects,” in IEEE Int. Conf. Robot. Autom. (ICRA) , 1998, pp. 2409–2415

    work page 1998

  7. [7]

    Optimization of power grasps for multiple objects,

    T. Yoshikawa, T. Watanabe, and M. Daito, “Optimization of power grasps for multiple objects,” in IEEE Int. Conf. Robot. Autom. (ICRA) , 2001, pp. 1786–1791

    work page 2001

  8. [8]

    Grasp stability analysis of multiple planar objects,

    T. Yamada, S. Yamanaka, M. Yamada, Y . Funahashi, and H. Yamamoto, “Grasp stability analysis of multiple planar objects,” in IEEE Int. Conf. Robot. Biomimetics , 2009, pp. 1032–1038

    work page 2009

Show all 29 references
  1. [9]

    Stability analysis of multiple objects grasped by multi-fingered hands with revolute joints in 2d,

    T. Yamada, M. Yamada, and H. Yamamoto, “Stability analysis of multiple objects grasped by multi-fingered hands with revolute joints in 2d,” in IEEE Int. Conf. Mechatronics Autom., 2012, pp. 1785–1792

    2012

  2. [10]

    Static grasp stability analysis of multiple spatial objects,

    T. Yamada and H. Yamamoto, “Static grasp stability analysis of multiple spatial objects,” J. Control Sci. Eng. , vol. 3, pp. 118–139, 2015

    2015

  3. [11]

    Multi-object grasping – estimating the number of ob- jects in a robotic grasp,

    T. Chen, A. Shenoy, A. Kolinko, S. Shah, and Y . Sun, “Multi-object grasping – estimating the number of ob- jects in a robotic grasp,” in IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS) , 2021, pp. 4995–5001

    2021

  4. [12]

    Multi-object grasping- efficient robotic picking and transferring policy for batch picking,

    A. Shenoy, T. Chen, and Y . Sun, “Multi-object grasping- efficient robotic picking and transferring policy for batch picking,” in IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), 2022, pp. 2741–2747

    2022

  5. [13]

    Uncertainty-aware self-supervised target-mass grasping of granular foods,

    K. Takahashi, W. Ko, A. Ummadisingu, and S.-i. Maeda, “Uncertainty-aware self-supervised target-mass grasping of granular foods,” in IEEE Int. Conf. Robot. Autom. (ICRA), 2021, pp. 2620–2626

    2021

  6. [14]

    Target- mass grasping of entangled food using pre-grasping & post-grasping,

    K. Takahashi, N. Fukaya, and A. Ummadisingu, “Target- mass grasping of entangled food using pre-grasping & post-grasping,” IEEE Robot. Autom. Lett. , vol. 7, no. 2, pp. 1222–1229, 2021

    2021

  7. [15]

    Grasp multiple objects with one hand,

    Y . Li, B. Liu, Y . Geng, P. Li, Y . Yang, Y . Zhu, T. Liu, and S. Huang, “Grasp multiple objects with one hand,” IEEE Robot. Autom. Lett. , vol. 9, no. 5, pp. 4027–4034, 2024

    2024

  8. [16]

    Multiple- object grasping using a multiple-suction-cup vacuum gripper in cluttered scenes,

    P. Jiang, J. Oaki, Y . Ishihara, and J. Ooga, “Multiple- object grasping using a multiple-suction-cup vacuum gripper in cluttered scenes,” Robotics, vol. 13, no. 6, 2024

    2024

  9. [17]

    Wiring-claw gripper for soft-stable picking up multiple objects,

    V . P. Nguyen and W. T. Chow, “Wiring-claw gripper for soft-stable picking up multiple objects,” IEEE Robot. Autom. Lett. , vol. 8, no. 7, pp. 3972–3979, 2023

    2023

  10. [18]

    Exploiting kinematic redundancy for robotic grasping of multiple objects,

    K. Yao and A. Billard, “Exploiting kinematic redundancy for robotic grasping of multiple objects,” IEEE Trans. Robot., vol. 39, no. 3, pp. 1982–2002, 2023

    1982

  11. [19]

    Learning to efficiently plan robust frictional multi-object grasps,

    W. C. Agboh, S. Sharma, K. Srinivas, M. Parulekar, G. Datta, T. Qiu, J. Ichnowski, E. Solowjow, M. Dogar, and K. Goldberg, “Learning to efficiently plan robust frictional multi-object grasps,” in IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS) , 2023, pp. 10 660–10 667

    2023

  12. [20]

    Push-mog: Efficient pushing to con- solidate polygonal objects for multi-object grasping,

    S. Aeron, E. LLontop, A. Adler, W. C. Agboh, M. Dogar, and K. Goldberg, “Push-mog: Efficient pushing to con- solidate polygonal objects for multi-object grasping,” in IEEE Int. Conf. Autom. Sci. Eng. (CASE) , 2023, pp. 1–6

    2023

  13. [21]

    Only pick once – multi-object picking algorithms for picking exact number of objects efficiently,

    Z. Ye and Y . Sun, “Only pick once – multi-object picking algorithms for picking exact number of objects efficiently,” arXiv:2307.02662, 2023

    2023 arXiv

  14. [22]

    The busboy problem: Efficient tableware decluttering using consolidation and multi-object grasps,

    K. Srinivas, S. Ganti, R. Parikh, A. Ahmad, W. Agboh, M. Dogar, and K. Goldberg, “The busboy problem: Efficient tableware decluttering using consolidation and multi-object grasps,” in IEEE Int. Conf. Autom. Sci. Eng. (CASE), 2023, pp. 1–6

    2023

  15. [23]

    Learning synergies between pushing and grasping with self-supervised deep reinforcement learning,

    A. Zeng, S. Song, S. Welker, J. Lee, A. Rodriguez, and T. Funkhouser, “Learning synergies between pushing and grasping with self-supervised deep reinforcement learning,” in IEEE/RSJ Int. Conf. Intell. Robots Syst. (IROS), 2018, pp. 4238–4245

    2018

  16. [24]

    Combining reinforcement learning and rule-based method to manipulate objects in clutter,

    Y . Chen, Z. Ju, and C. Yang, “Combining reinforcement learning and rule-based method to manipulate objects in clutter,” in Int. Joint Conf. Neural Netw. (IJCNN) , 2020, pp. 1–6

    2020

  17. [25]

    Self-supervised learning for joint pushing and grasping policies in highly cluttered environments,

    Y . Wang, K. Mokhtar, C. Heemskerk, and H. Kasaei, “Self-supervised learning for joint pushing and grasping policies in highly cluttered environments,” in IEEE Int. Conf. Robot. Autom. (ICRA) , 2024, pp. 13 840–13 847

    2024

  18. [26]

    Learning fine-grained bimanual manipulation with low-cost hard- ware,

    T. Z. Zhao, V . Kumar, S. Levine, and C. Finn, “Learning fine-grained bimanual manipulation with low-cost hard- ware,” arXiv:2304.13705, 2023

    2023 arXiv

  19. [27]

    Anyteleop: A gen- eral vision-based dexterous robot arm-hand teleoperation system,

    Y . Qin, W. Yang, B. Huang, K. Van Wyk, H. Su, X. Wang, Y .-W. Chao, and D. Fox, “Anyteleop: A gen- eral vision-based dexterous robot arm-hand teleoperation system,” in Robot.: Sci. Syst. , 2023

    2023

  20. [28]

    Wilor: End-to-end 3d hand localization and reconstruc- tion in-the-wild,

    R. A. Potamias, J. Zhang, J. Deng, and S. Zafeiriou, “Wilor: End-to-end 3d hand localization and reconstruc- tion in-the-wild,” in IEEE Conf. Comput. Vis. Pattern Recognit., 2025

    2025

  21. [29]

    Benchmarking multi-object grasping,

    T. Chen, R. Frumento, G. Pagnanelli, G. Cei, V . Keth, S. Gafarov, J. Gong, Z. Ye, M. Baracca, S. D’Avella, M. Bianchi, and Y . Sun, “Benchmarking multi-object grasping,” arXiv:2503.20820, 2025

    2025 arXiv

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.