Pith. sign in

REVIEW 4 major objections 5 minor 46 references

GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object Manipulation

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read GAPartManip claims that a synthetic dataset with material-randomized IR rendering and part-level interaction pose labels enables zero-shot sim-to-real manipulation of articulated objects, achieving 61.29% real-world success.

desk verdict Useful new dataset and a plausible framework for material-agnostic articulated object manipulation, but the real-world evaluation is too thin to carry the sim-to-real claim as it stands. read the letter →

arxiv 2411.18276 v3 pith:FBW3DC5U submitted 2024-11-27 cs.RO cs.AI

Wenbo Cui , Chengyang Zhao , Songlin Wei , Jiazhao Zhang , Haoran Geng , Yaran Chen , Haoran Li , He Wang This is my paper · ORCID
classification cs.ROcs.AI
keywords articulatedobjectmanipulationsim-to-realtransferdepthestimationactionableinteractionposestructuredlightcameradomainrandomizationpart-centricdatasetmaterial-agnosticperception
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

GAPartManip is a synthetic dataset built to test a data-centric hypothesis: that the gap between simulation and real-world manipulation of articulated objects is mostly a depth-sensing gap, and that it can be closed by training on infrared images that reproduce how structured-light cameras fail on transparent and reflective materials. The paper reports that fine-tuning a diffusion-based depth estimator and a part-aware grasp pose predictor on this dataset sharply reduces depth error in simulation and raises real-world open/close manipulation success to 61.29%, compared with 32.25% for the best baseline. The claim is not a new algorithm but a new data resource: 918 objects from 19 household categories, 240K rendered images, and over 8 billion scene-level actionable interaction poses.

What carries the argument

The load-bearing mechanism is the material-randomized IR rendering pipeline built on NVIDIA Isaac Sim, which replicates the Intel RealSense D415's structured-light imaging layout (IR projector, two IR cameras, projected shadow pattern) and randomizes diffuse, transparent, specular, and metal material parameters, ambient lighting, backgrounds, and joint states across 20 indoor scenes. The second mechanism is the CUDA-accelerated scene-level pose annotation pipeline, which projects part-level antipodal grasps into the scene and filters them by single-view visibility and collision, reducing annotation time from five minutes to under two seconds per part and producing over 8 billion actionable poses. Together these two mechanisms turn a simulator into a training distribution that is deliberately material-agnostic, so models trained on it do not assume clean depth or reflective surfaces.

What would settle it

Collect real Intel RealSense D415 IR and depth captures of transparent-lidded and reflective-handled articulated objects under lighting conditions not represented in the 20 simulated scenes, and measure the depth error of the GAPartManip-trained depth model against the raw sensor output; if the model is not consistently better than raw depth on these materials, the material-randomized IR simulation is not capturing the sensor's real failure modes.

Watch

Extended reading notes

Core claim

The central claim is that a large-scale part-centric dataset with physics-based IR rendering and scene-level actionable interaction pose annotations is sufficient to make articulated object manipulation work zero-shot in the real world, even when object parts are transparent or reflective. On the depth side, fine-tuning D3RoMa on GAPartManip improves simulated MAE from 0.317 to 0.050 and EPE from 2.82 to 0.69, and qualitative real-world captures show cleaner depth on transparent lids and small handles. On the pose side, the proposed Part-aware EcoGrasp, trained on the dataset's actioness labels, reaches 55.33% precision on seen instances and 41.65% on novel instances, versus 24.72% and 14.56% for the untrained EconomicGrasp baseline. Combined with a motion-direction estimator and a local planner, the full framework achieves 61.29% success across 31 real-world open/close trials, against 32.25% for the best baseline.

Load-bearing premise

The result depends on the simulator's modeled failures of the Intel RealSense D415 on transparent and reflective surfaces being close enough to the real camera's failures that models trained on synthetic IR images transfer to real IR images.

Editorial extensions

If this is right

  • Fine-tuning depth networks on GAPartManip cuts simulated mean depth error roughly sixfold (MAE from 0.317 to 0.050), a direct consequence of the material-randomized IR data.
  • Pose predictors trained on the dataset learn to score parts by actionability rather than graspability alone, so predicted poses open and close articulated objects instead of merely holding them; the paper reports 55.33% precision on seen instances versus 24.72% for the same architecture without dataset training.
  • The full pipeline transfers zero-shot to a real robot arm and Intel RealSense D415, reaching 61.29% success on 31 open/close trials, more than the 32.25% of the best baseline.
  • Because the dataset covers 19 categories with seen, similar, and novel instance splits, models trained on it generalize to object instances not used in training.

Reading between the lines

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

  • Editorial inference: if the D415 simulation is faithful, the same material-randomized IR pipeline could transfer to other structured-light depth sensors, because the domain randomization targets sensor physics rather than category-specific appearance; this is testable by fine-tuning the same models on synthetic IR from a different camera model.
  • Editorial inference: the reported real-world advantage rests on 31 trials across 7 object instances, so a larger multi-scene evaluation would be needed to confirm the magnitude of the gap over baselines, which the paper does not provide.
  • Editorial inference: the pose annotations are designed for a single parallel-jaw gripper with antipodal scoring, so extending the dataset to suction or multi-finger grippers would require re-annotation, though the part-level actionable pose representation is a plausible starting point.
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

4 major / 5 minor

Summary. The paper introduces GAPartManip, a large synthetic dataset for articulated object manipulation containing 918 object instances across 19 categories, 241,680 rendered RGB/IR/depth samples, and 8 billion scene-level actionable interaction pose annotations. The data generation pipeline combines physics-based simulation of an Intel RealSense D415 structured-light camera with domain-randomized materials and a CUDA-accelerated pose annotation procedure. The authors also propose a modular manipulation framework comprising a D3RoMa-based depth reconstruction module, a Part-aware EcoGrasp pose prediction module, and a CuRobo-based local planner. Evaluations report that fine-tuning depth and pose models on GAPartManip substantially improves simulated depth metrics (Table I) and simulated actionable-pose precision (Table II), and that the full framework achieves a 61.29% real-world open/close success rate over 31 trials versus 32.25% for the best baseline (Table III).

Significance. If the reported results hold up, GAPartManip would be a valuable resource for the community: it is the first articulated-object manipulation dataset to combine material-randomized IR rendering with dense scene-level actionable pose annotations, and the proposed framework demonstrates a promising route to zero-shot sim-to-real manipulation of transparent and reflective parts. The paper also contributes a substantial engineering effort in the CUDA-accelerated pose filtering pipeline (about 150x speedup) and a modular system where each component is evaluated separately and in ablation. The main external validation is the real-world manipulation experiment, which compares against several existing methods. However, the strength of this contribution is currently limited by the thin real-world evidence and the lack of quantitative validation of the simulated D415 sensor model.

major comments (4)
  1. [Section V-A and Fig. 5] The real-world depth evaluation is qualitative only. The paper states that the fine-tuned models "demonstrate strong depth estimation performance in real-world scenarios" and that "our model significantly outperforms the fine-tuned RAFT-Stereo and the raw depth," but Figure 5 shows no quantitative comparison (no ground-truth depth, no error metric) on real D415 data. Since the depth module is load-bearing (Table III: removing it drops overall success from 61.29% to 38.70%), the simulated depth gains in Table I (e.g., MAE 0.317 to 0.050) are not confirmed on real sensor data. Please provide quantitative real-world depth evaluation, for example by comparing against a high-accuracy depth sensor or manual ground truth on a small set of real scenes, or at least report per-scene depth errors for the cases shown in Figure 5.
  2. [Section III-B] The fidelity of the simulated D415 structured-light camera is asserted but not validated. Section III-B states that the authors "replicate the layout of the D415 imaging system" and "project a similar shadow pattern," and that material parameters are randomized, but no experiment demonstrates that the resulting simulated IR images and depth corruptions quantitatively match a real D415 on transparent, specular, or metal surfaces. The simulation results in Table I are measured on test images produced by the same rendering pipeline that generated the training data, so they do not by themselves establish sim-to-real transfer. A concrete validation would be to capture real D415 IR and raw depth for a small set of objects (or a re-created scene) and compare the simulated and real sensor outputs, or to measure real-world depth reconstruction error against a reference sensor before and after fine-tuning.
  3. [Section V-C and Table III] The real-world manipulation results are based on only 31 total trials (14 open, 17 close) and are reported without error bars, confidence intervals, or a significance test. With n=31, the 95% confidence interval for a 61.29% success rate is roughly [42%, 78%], and for the 32.25% baseline it is roughly [17%, 51%]; the intervals overlap. The claim of a "significant performance boost" relative to baselines is therefore not statistically supported. Please report per-instance trial counts, confidence intervals, or a two-proportion test, and if possible increase the number of trials, especially for the Ours vs. GSNet comparison.
  4. [Section V-C] The baseline comparison in the real-world experiment is underspecified. The text says "we compare our method with three baselines: Where2Act, AO-Grasp, GSNet, and, like in Sec. V-B, we extend the Where2Act interaction pipeline to finish our tasks." It is not clear whether AO-Grasp and GSNet are also given the same post-grasping motion direction estimator, the same CuRobo planner, and the same mechanism for executing open/close actions, or whether they only provide a grasp pose with no articulated-object interaction. If the baselines are not integrated with the same interaction pipeline, the comparison may be unfair. Please describe the exact setup for each baseline, including how the open/close action is determined after grasping.
minor comments (5)
  1. [Section V-A] The percentage improvements in the text are incorrect. For RAFT-Stereo, MAE goes from 0.618 to 0.309, which is a 50% error reduction (a 100% improvement, not 150%). For D3RoMa, MAE goes from 0.317 to 0.050, which is an 84% error reduction (about 534% improvement, not 600%). Please correct the percentages.
  2. [Figure 4] The word "TODO" appears in the framework diagram. This looks like an unfinished annotation and should be removed before publication.
  3. [Section II-A and II-B] There are two citation placeholders marked "[?]" in the related work. These should be filled in with the intended references.
  4. [Table II] The column header for the Novel split reads "P P 0.4 P0.8" while the Seen and Unseen splits read "P P 0.8 P0.4". If this is not a deliberate reordering, it is a typo and should be corrected.
  5. [Section V-C] The paper states the real-world experiments use "7 distinct instances, including StorageFurniture, Box, and Microwave" but does not report how many trials were conducted per instance or per object. Providing a per-instance breakdown would help readers assess where the method succeeds and fails.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: real-world manipulation benchmarks provide external validation; simulated metrics are standard in-distribution dataset evaluations.

full rationale

GAPartManip's headline claims are empirical dataset-utility claims, and the load-bearing evidence is external rather than definitional. The simulated depth (Table I) and pose (Table II) numbers are held-out evaluations on the same Isaac Sim rendering/annotation pipeline that produced the training set, so they measure in-distribution learning; this is a standard way to demonstrate a dataset's usefulness, not a case of a fitted parameter being renamed as a prediction. The sim-to-real claim is supported by independent real-world manipulation trials (Table III: 61.29% vs. 32.25% best baseline over 31 trials) and by ablations showing the depth and pose modules each contribute (dropping depth reconstruction reduces success to 38.70%; dropping Part-aware EcoGrasp to 51.61%). Self-citations to GAPartNet [9] provide object assets, part semantics, and a motion prior, and D3RoMa [14] provides the depth architecture, but no equation in the paper defines the reported improvements as the dataset's own labels or forces the outcome; the improvements are measured against external baselines and a real robot. The small real-world trial count and the absence of quantitative real-depth ground truth are validity limitations, not circularity.

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

The ledger is dominated by domain assumptions rather than fitted physical parameters. The paper hand-sets sampling densities and thresholds, assumes GAPartNet semantics, and assumes that the simulated D415 IR response and grasp simulation are faithful proxies. These assumptions are load-bearing for both labels and evaluation, so external validation of the dataset is currently limited.

free parameters (4)
  • Pose sampling grid (N, V, A, D) = N=512, V=64, A=12, D=4
    Hand-chosen sampling density for FPS candidate points, gripper views, in-plane rotations, and depths; determines annotation count and coverage.
  • Actionability threshold T = not reported numerically
    Threshold used to filter low-quality poses in Eq. (1) and Eq. (2); value omitted, though performance depends on it.
  • Domain randomization ranges = ambient lighting, background, diffuse/transparent/specular/metal material parameters, camera latitude/longitude ranges
    Hand-set randomization bounds intended to close the sim-to-real gap; no sensitivity analysis is provided.
  • Friction coefficients for pose evaluation = P, P0.8, P0.4 (unspecified values)
    Used to define grasp success in the precision metric; chosen by convention from GraspNet evaluation.
assumptions (4)
  • domain assumption Synthetic IR rendering with randomized materials faithfully approximates real Intel RealSense D415 imaging behavior for transparent and reflective surfaces.
    The central sim-to-real depth claim depends on this fidelity; Section III-B describes the D415 replication but provides no quantitative validation of the sensor model beyond qualitative real-world images.
  • domain assumption GAPartNet semantic part annotations and part poses are correct and complete for all 918 instances.
    All actionability labels and motion-direction predictions inherit from GAPartNet [9]; errors there propagate into the dataset labels.
  • domain assumption Antipodal analysis and geometric collision filtering predict physically successful grasps.
    Both the generated pose labels and the simulation precision metric rely on antipodal grasp-quality scoring from GraspNet; this is a standard proxy, not a physical grasp guarantee (Sections III-C, V-B).
  • domain assumption The simulation environment used for precision evaluation is a valid proxy for real-world manipulation success.
    Simulation precision numbers are reported without a calibrated mapping to real trials; real-world support rests on 31 trials.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object Manipulation." pith.science (2026). https://pith.science/paper/FBW3DC5U

@misc{pith2026241118276,
  author       = {Pith},
  title        = {Pith review of: GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object Manipulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FBW3DC5U}},
  note         = {Machine review of arXiv:2411.18276}
}
read the original abstract

Effectively manipulating articulated objects in household scenarios is a crucial step toward achieving general embodied artificial intelligence. Mainstream research in 3D vision has primarily focused on manipulation through depth perception and pose detection. However, in real-world environments, these methods often face challenges due to imperfect depth perception, such as with transparent lids and reflective handles. Moreover, they generally lack the diversity in part-based interactions required for flexible and adaptable manipulation. To address these challenges, we introduced a large-scale part-centric dataset for articulated object manipulation that features both photo-realistic material randomization and detailed annotations of part-oriented, scene-level actionable interaction poses. We evaluated the effectiveness of our dataset by integrating it with several state-of-the-art methods for depth estimation and interaction pose prediction. Additionally, we proposed a novel modular framework that delivers superior and robust performance for generalizable articulated object manipulation. Our extensive experiments demonstrate that our dataset significantly improves the performance of depth perception and actionable interaction pose prediction in both simulation and real-world scenarios. More information and demos can be found at: https://pku-epic.github.io/GAPartManip/.

Figures

Figures reproduced from arXiv: 2411.18276 by the authors.

Figure 1
Figure 1. GAPartManip. We introduce a large-scale part-centric dataset for material-agnostic articulated object manipulation. It encompasses 19 common household articulated categories, totaling 918 object instances, 240K photo-realistic rendering images, and 8 billion scene-level actionable interaction poses. GAPartManip enables robust zero-shot sim-to-real transfer for accomplishing articulated object manipulation tasks. Fir… view at source ↗
Figure 2
Figure 2. Data Examples in GAPartManip. GAPartManip is a novel large-scale synthetic dataset for articulated objects, featuring two important aspects: 1) realistic, physics-based IR rendering for various object materials in diverse scenes, and 2) part-oriented actionable interaction pose annotations for a wide range of articulated objects. Each column shows a data sample. From top to bottom, each row displays the RGB image, t… view at source ↗
Figure 3
Figure 3. Dataset Generation Pipeline. For scene-level data sample rendering, we input the object asset into our photo-realistic rendering pipeline, generating one RGB image and two IR images (left and right) for each camera perspective. For pose annotation, we begin by performing mesh fusion on each GAPart on the object to establish a one-to-one correspondence between GAParts and meshes. Then, we use FPS to obtain the point … view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Framework Overview. Given IR images and raw depth map, the depth reconstruction module first performs depth recovery. Subsequently, the pose prediction module generates a 7-DoF actionable pose and a 6-DoF post-grasping motion for interaction based on the reconstructed …
Figure 5
Figure 5. Figure 5: Qualitative Results for Depth Estimation in the Real World. Our refined depth maps are cleaner and more accurate than the ones from the baseline, indicating that our depth reconstruction module is more robust for transparent and translucent lids and small handles. Zoom…
Figure 6
Figure 6. Figure 6: Qualitative Comparison of Actionable Pose Prediction in Simulation. estimation. Data Preparation. We split the dataset into training and testing sets using an approximate 7:3 ratio. We further divide the testing set into 3 categories: seen instances, unseen but similar…
Figure 7
Figure 7. Figure 7: Qualitative Results For Real-world Manipulation. The actionable poses with top scores are displayed, with the red gripper representing the top-1 pose. parts rather than the non-actionable ones. For instance, although the pre-trained EconomicGrasp in [PITH_FULL_IMAGE:f…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 34 canonical work pages

  1. [1]

    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,” in The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2020. 1, 2

    work page 2020

  2. [2]

    Where2act: From pixels to actions for articulated 3d objects,

    K. Mo, L. J. Guibas, M. Mukadam, A. Gupta, and S. Tulsiani, “Where2act: From pixels to actions for articulated 3d objects,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 6813–6823. 1, 2, 6, 7

    work page 2021

  3. [3]

    Vat-mart: Learning visual action trajectory proposals for manipulating 3d articulated objects,

    R. Wu, Y . Zhao, K. Mo, Z. Guo, Y . Wang, T. Wu, Q. Fan, X. Chen, L. Guibas, and H. Dong, “Vat-mart: Learning visual action trajectory proposals for manipulating 3d articulated objects,” arXiv preprint arXiv:2106.14440, 2021. 1, 2

    arXiv 2021

  4. [4]

    Adaafford: Learning to adapt manipulation affordance for 3d artic- ulated objects via few-shot interactions,

    Y . Wang, R. Wu, K. Mo, J. Ke, Q. Fan, L. J. Guibas, and H. Dong, “Adaafford: Learning to adapt manipulation affordance for 3d artic- ulated objects via few-shot interactions,” in European conference on computer vision . Springer, 2022, pp. 90–107. 1

    work page 2022

  5. [5]

    Dualafford: Learning collaborative visual affordance for dual-gripper manipulation,

    Y . Zhao, R. Wu, Z. Chen, Y . Zhang, Q. Fan, K. Mo, and H. Dong, “Dualafford: Learning collaborative visual affordance for dual-gripper manipulation,” arXiv preprint arXiv:2207.01971 , 2022. 1

    arXiv 2022

  6. [6]

    Flowbot3d: Learning 3d ar- ticulation flow to manipulate articulated objects,

    B. Eisner, H. Zhang, and D. Held, “Flowbot3d: Learning 3d ar- ticulation flow to manipulate articulated objects,” arXiv preprint arXiv:2205.04382, 2022. 1

    arXiv 2022

  7. [7]

    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,” arXiv preprint arXiv:2306.12893, 2023. 1

    arXiv 2023

  8. [8]

    3d implicit transporter for temporally consistent keypoint discovery,

    C. Zhong, Y . Zheng, Y . Zheng, H. Zhao, L. Yi, X. Mu, L. Wang, P. Li, G. Zhou, C. Yang, et al., “3d implicit transporter for temporally consistent keypoint discovery,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023, pp. 3869–3880. 1

    work page 2023

Show all 46 references
  1. [9]

    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,” in Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 202...

    2023

  2. [10]

    Partmanip: Learning cross-category generalizable part manipulation policy from point cloud observations,

    H. Geng, Z. Li, Y . Geng, J. Chen, H. Dong, and H. Wang, “Partmanip: Learning cross-category generalizable part manipulation policy from point cloud observations,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2023, pp. 2978–2988. 1, 2

    2023

  3. [11]

    Sage: Bridging semantic and actionable parts for generalizable manipulation of articulated objects,

    H. Geng, S. Wei, C. Deng, B. Shen, H. Wang, and L. Guibas, “Sage: Bridging semantic and actionable parts for generalizable manipulation of articulated objects,” 2024. 1, 2

    2024

  4. [12]

    Rpmart: Towards robust perception and manipulation for articulated objects,

    J. Wang, W. Liu, Q. Yu, Y . You, L. Liu, W. Wang, and C. Lu, “Rpmart: Towards robust perception and manipulation for articulated objects,” arXiv preprint arXiv:2403.16023 , 2024. 1, 2, 6

    2024 arXiv

  5. [13]

    Rlafford: End-to-end affordance learning for robotic manipulation,

    Y . Geng, B. An, H. Geng, Y . Chen, Y . Yang, and H. Dong, “Rlafford: End-to-end affordance learning for robotic manipulation,” in 2023 IEEE International Conference on Robotics and Automation (ICRA) , 2023, pp. 5880–5886. 1, 2

    2023

  6. [14]

    D3roma: Disparity diffusion-based depth sensing for material-agnostic robotic manipulation,

    S. Wei, H. Geng, J. Chen, C. Deng, W. Cui, C. Zhao, X. Fang, L. Guibas, and H. Wang, “D3roma: Disparity diffusion-based depth sensing for material-agnostic robotic manipulation,” in 8th Annual Conference on Robot Learning (CoRL) , 2024. 1, 3, 4, 5

    2024

  7. [15]

    Asgrasp: Generalizable transparent object reconstruction and 6-dof grasp detection from rgb-d active stereo camera,

    J. Shi, A. Yong, Y . Jin, D. Li, H. Niu, Z. Jin, and H. Wang, “Asgrasp: Generalizable transparent object reconstruction and 6-dof grasp detection from rgb-d active stereo camera,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 5441–5447. 1, 2

    2024

  8. [16]

    Ao-grasp: Articulated object grasp generation,

    C. P. Morlans, C. Chen, Y . Weng, M. Yi, Y . Huang, N. Heppert, L. Zhou, L. Guibas, and J. Bohg, “Ao-grasp: Articulated object grasp generation,” arXiv preprint arXiv:2310.15928 , 2023. 2, 7

    2023 arXiv

  9. [17]

    Rgbmanip: Monocular image-based robotic manipulation through active object pose estimation,

    B. An, Y . Geng, K. Chen, X. Li, Q. Dou, and H. Dong, “Rgbmanip: Monocular image-based robotic manipulation through active object pose estimation,” in 2024 IEEE International Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 7748–7755. 2

    2024

  10. [18]

    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,” arXiv preprint arXiv:1809.07417, 2018. 2

    2018 arXiv

  11. [19]

    Banana: Banach fixed-point network for pointcloud segmentation with inter- part equivariance,

    C. Deng, J. Lei, W. B. Shen, K. Daniilidis, and L. J. Guibas, “Banana: Banach fixed-point network for pointcloud segmentation with inter- part equivariance,” in NeurIPS, 2024. 2

    2024

  12. [20]

    Category-level articulated object pose estimation,

    X. Li, H. Wang, L. Yi, L. J. Guibas, A. L. Abbott, and S. Song, “Category-level articulated object pose estimation,” in CVPR, 2020. 2

    2020

  13. [21]

    Semi-weakly supervised object kinematic motion prediction,

    G. Liu, Q. Sun, H. Huang, C. Ma, Y . Guo, L. Yi, H. Huang, and R. Hu, “Semi-weakly supervised object kinematic motion prediction,” in CVPR, 2023. 2

    2023

  14. [22]

    Scissorbot: Learning generalizable scissor skill for paper cutting via simulation, imitation, and sim2real,

    J. Lyu, Y . Chen, T. Du, F. Zhu, H. Liu, Y . Wang, and H. Wang, “Scissorbot: Learning generalizable scissor skill for paper cutting via simulation, imitation, and sim2real,” in 8th Annual Conference on Robot Learning , 2024. [Online]. Available: https://openreview.net/forum?id...

    2024

  15. [23]

    Gamma: Graspability-aware mobile manipulation policy learning based on online grasping pose fusion,

    J. Zhang, N. Gireesh, J. Wang, X. Fang, C. Xu, W. Chen, L. Dai, and H. Wang, “Gamma: Graspability-aware mobile manipulation policy learning based on online grasping pose fusion,” in 2024 IEEE Inter- national Conference on Robotics and Automation (ICRA) . IEEE, 2024, pp. 1399–1405. 2

    2024

  16. [24]

    Urdformer: Constructing interactive realistic scenes from real im- ages via simulation and generative modeling,

    Q. Chen, M. Memmel, A. Fang, A. Walsman, D. Fox, and A. Gupta, “Urdformer: Constructing interactive realistic scenes from real im- ages via simulation and generative modeling,” in Towards Generalist Robots: Learning Paradigms for Scalable Skill Acquisition @ CoRL 2023, 2023. 2

    2023

  17. [25]

    A-sdf: Learning disentangled signed distance functions for articulated shape representation,

    J. Mu, W. Qiu, A. Kortylewski, A. Yuille, N. Vasconcelos, and X. Wang, “A-sdf: Learning disentangled signed distance functions for articulated shape representation,” in ICCV, 2021. 2

    2021

  18. [26]

    Ditto: Building digital twins of articulated objects from interaction,

    Z. Jiang, C.-C. Hsu, and Y . Zhu, “Ditto: Building digital twins of articulated objects from interaction,” in CVPR, 2022. 2

    2022

  19. [27]

    Cla-nerf: Category-level articulated neural radiance field,

    W.-C. Tseng, H.-J. Liao, L. Yen-Chen, and M. Sun, “Cla-nerf: Category-level articulated neural radiance field,” in ICRA, 2022. 2

    2022

  20. [28]

    Physpart: Physically plausible part completion for interactable objects,

    R. Luo, H. Geng, C. Deng, P. Li, Z. Wang, B. Jia, L. Guibas, and S. Huang, “Physpart: Physically plausible part completion for interactable objects,” 2024. [Online]. Available: https://arxiv.org/abs/ 2408.13724 2

    2024 arXiv

  21. [29]

    Nap: Neural 3d articulation prior,

    J. Lei, C. Deng, B. Shen, L. Guibas, and K. Daniilidis, “Nap: Neural 3d articulation prior,” arXiv preprint arXiv:2305.16315 , 2023. 2

    2023 arXiv

  22. [30]

    Cage: Controllable articulation generation,

    J. Liu, H. I. I. Tam, A. Mahdavi-Amiri, and M. Savva, “Cage: Controllable articulation generation,” in CVPR, 2024. 2

    2024

  23. [31]

    End- to-end affordance learning for robotic manipulation,

    Y . Geng, B. An, H. Geng, Y . Chen, Y . Yang, and H. Dong, “End- to-end affordance learning for robotic manipulation,” in ICRA, 2023. 2

    2023

  24. [32]

    Arnold: A benchmark for language- grounded task learning with continuous states in realistic 3d scenes,

    R. Gong, J. Huang, Y . Zhao, H. Geng, X. Gao, Q. Wu, W. Ai, Z. Zhou, D. Terzopoulos, S.-C. Zhu, et al., “Arnold: A benchmark for language- grounded task learning with continuous states in realistic 3d scenes,” in ICCV, 2023. 2

    2023

  25. [33]

    Graspnet-1billion: A large- scale benchmark for general object grasping,

    H.-S. Fang, C. Wang, M. Gou, and C. Lu, “Graspnet-1billion: A large- scale benchmark for general object grasping,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2020, pp. 11 444–11 453. 2, 3, 4, 6

    2020

  26. [34]

    Contact- graspnet: Efficient 6-dof grasp generation in cluttered scenes,

    M. Sundermeyer, A. Mousavian, R. Triebel, and D. Fox, “Contact- graspnet: Efficient 6-dof grasp generation in cluttered scenes,” 2021. [Online]. Available: https://arxiv.org/abs/2103.14127 2

    2021 arXiv

  27. [35]

    Instruction-driven history-aware policies for robotic ma- nipulations,

    P.-L. Guhur, S. Chen, R. G. Pinel, M. Tapaswi, I. Laptev, and C. Schmid, “Instruction-driven history-aware policies for robotic ma- nipulations,” in Conference on Robot Learning . PMLR, 2023, pp. 175–187. 2

    2023

  28. [36]

    Composable part-based manipulation,

    W. Liu, J. Mao, J. Hsu, T. Hermans, A. Garg, and J. Wu, “Composable part-based manipulation,” 2024. [Online]. Available: https://arxiv.org/abs/2405.05876 2

    2024 arXiv

  29. [37]

    Articulated object manipulation with coarse-to-fine affordance for mitigating the effect of point cloud noise,

    S. Ling, Y . Wang, S. Wu, Y . Zhuang, T. Xu, Y . Li, C. Liu, and H. Dong, “Articulated object manipulation with coarse-to-fine affordance for mitigating the effect of point cloud noise,” 2024. [Online]. Available: https://arxiv.org/abs/2402.18699 2

    2024 arXiv

  30. [38]

    Gpu-accelerated robotic simulation for distributed reinforce- ment learning,

    J. Liang, V . Makoviychuk, A. Handa, N. Chentanez, M. Macklin, and D. Fox, “Gpu-accelerated robotic simulation for distributed reinforce- ment learning,” 2018. 3

    2018

  31. [39]

    Domain randomization-enhanced depth simulation and restoration for perceiving and grasping specular and transparent objects,

    Q. Dai, J. Zhang, Q. Li, T. Wu, H. Dong, Z. Liu, P. Tan, and H. Wang, “Domain randomization-enhanced depth simulation and restoration for perceiving and grasping specular and transparent objects,” in European Conference on Computer Vision (ECCV) , 2022. 3

    2022

  32. [40]

    An economic framework for 6-dof grasp detection,

    X.-M. Wu, J.-F. Cai, J.-J. Jiang, D. Zheng, Y .-L. Wei, and W.-S. Zheng, “An economic framework for 6-dof grasp detection,” 2024. [Online]. Available: https://arxiv.org/abs/2407.08366 4, 6

    2024 arXiv

  33. [41]

    Curobo: Parallelized collision-free robot motion gen- eration,

    B. Sundaralingam, S. K. S. Hari, A. Fishman, C. Garrett, K. Van Wyk, V . Blukis, A. Millane, H. Oleynikova, A. Handa, F. Ramos, N. Ratliff, and D. Fox, “Curobo: Parallelized collision-free robot motion gen- eration,” in 2023 IEEE International Conference on Robotics and Automa...

    2023

  34. [42]

    Stereo processing by semiglobal matching and mutual information,

    H. Hirschmuller, “Stereo processing by semiglobal matching and mutual information,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 30, no. 2, pp. 328–341, 2008. 5

    2008

  35. [43]

    Raft-stereo: Multilevel recurrent field transforms for stereo matching,

    L. Lipson, Z. Teed, and J. Deng, “Raft-stereo: Multilevel recurrent field transforms for stereo matching,” in 2021 International Conference on 3D Vision (3DV) . IEEE, 2021, pp. 218–227. 5

    2021

  36. [44]

    Raft: Recurrent all-pairs field transforms for optical flow,

    Z. Teed and J. Deng, “Raft: Recurrent all-pairs field transforms for optical flow,” in Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part II

    2020

  37. [45]

    Springer, 2020, pp. 402–419. 5

    2020

  38. [46]

    Grasp- ness discovery in clutters for fast and accurate grasp detection,

    C. Wang, H.-S. Fang, M. Gou, H. Fang, J. Gao, and C. Lu, “Grasp- ness discovery in clutters for fast and accurate grasp detection,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 15 964–15 973. 6, 7

    2021

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.