Pith. sign in

REVIEW 5 major objections 6 minor 43 references

A Multi-Level Similarity Approach for Single-View Object Grasping: Matching, Planning, and Fine-Tuning

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a robot can grasp never-seen objects from a single camera view by matching the observed point cloud to a small database of known models and transferring their preplanned grasps, without any training.

desk verdict Inventive training-free grasping pipeline, but the headline 'unknown-object' result is undermined because test objects likely sit in the matching database; the paper needs a clean hold-out evaluation. read the letter →

arxiv 2507.11938 v1 pith:BPVKFJFW submitted 2025-07-16 cs.RO cs.CV

Hao Chen , Takuya Kiyokawa , Zhengtao Hu , Weiwei Wan , Kensuke Harada This is my paper · ORCID
classification cs.ROcs.CV MSC 68T4068T45
keywords single-viewobjectgraspingsimilaritymatchingC-FPFHdescriptorgraspplanningpointcloudregistrationfine-tuninglearning-freeroboticsmulti-level
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that single-view grasping of unknown objects does not require a learned grasp model. It proposes a training-free pipeline that first matches the observed partial point cloud against a database of fewer than 100 known 3D models using three independent similarity levels, then transfers preplanned grasps from the best-matching models and refines them for stability. The central technical novelty is the C-FPFH descriptor, which lets partial and complete point clouds of different objects be compared. The authors argue that similarity matching is naturally robust to sensing noise and environmental change, and that is why their real-robot results beat learning-based benchmarks in isolated, cluttered, and adversarial scenes.

What carries the argument

The load-bearing object is the C-FPFH descriptor, a clustered Fast Point Feature Histogram. It reduces each sampled point's 33-dimensional FPFH vector to its two most dominant components, counts how often each such feature pair occurs across the cloud, and groups points sharing a feature pair into spatially coherent clusters whose PCA singular values describe local shape. This construction lets a partial single-view cloud be compared with a complete database cloud through QS, the fraction of the target's feature-pair mass present in the candidate, and DS, the distance between the target's main cluster shape and the closest similar cluster in the candidate. Around it sit the multi-level selection rule, the semi-oriented bounding box for dimensional matching, and plane-detection registration that replaces RANSAC in the coarse alignment step.

What would settle it

Run the isolated-object protocol on a set of objects whose 3D models are provably absent from the database and check both the identity of the matched reference and per-object success. If nonidentical matched models still yield near-96% success, the generalization claim is confirmed; if success drops or the pipeline only succeeds when the exact model sits in the database, the claim is not.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that object similarity can be decomposed into three independently evaluated dimensions and used as a reliable substitute for grasp prediction. The system computes a category-based semantic match with help from a large language model, a geometry match using the new C-FPFH descriptor and its two metrics (QS for feature-pair overlap and DS for distributional similarity), and a size match using a semi-oriented bounding box whose vertical axis is fixed by the supporting plane. Candidate models that pass these filters are ranked by plane-detection-based point cloud registration, their preplanned grasps are transformed onto the target cloud, and a two-stage fine-tuning step adjusts position and center based on local normals. The paper reports 96% grasp success and 98% plan success on ten isolated objects, around 90% success in cluttered scenes, and consistent performance in table-top, on-shelf, and hand-over environments, with lower planning time than the learning-based baselines.

Load-bearing premise

The load-bearing premise is that the test objects are truly unknown to the database; Section IV-B says the opposite for the matching experiment, because the authors add each object's original 3D model to the database, so the 96% success could be exact-model retrieval rather than similarity-based generalization to unseen shapes.

Editorial extensions

If this is right

  • A robot can acquire the ability to grasp a new category by inserting one or few 3D models with preplanned grasps into the database, with no training run.
  • The geometric matching level becomes the critical component under occlusion, while semantic and dimensional levels mainly narrow the candidate set; the paper's ablation shows removing geometric matching hurts most in occluded scenes.
  • Grasp fine-tuning contributes roughly as much as multi-level matching in the paper's ablation, increasing success by over 20% in their trials when triggered.
  • The method's planning time stays near 2 seconds with a small database and rises only about 1 second when the database grows to over 1,000 models through semantic and dimensional pre-filtering.

Reading between the lines

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

  • A decisive held-out test would remove the tested objects' original models from the database before matching; the matching experiment explicitly adds those models, so part of the reported accuracy may currently be exact-model retrieval rather than transfer across different shapes.
  • The ceiling of the approach is set by database coverage: an object with no geometrically similar neighbor should degrade toward the 'potential grasp' fallback, and the paper's appendix already notes higher failure risk when matches have low partial similarity.
  • Fixing the number of dominant FPFH components at two is a heuristic; the paper itself observes that the dominance gap shrinks on complex surfaces, suggesting an adaptive component count as a natural extension.
  • The large-database experiment pre-filters with Word2Vec before LLM matching, hinting at a scalable two-tier architecture: cheap embedding filters first, then expensive geometric comparison only on survivors.
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. The paper proposes a training-free, three-stage pipeline for single-view object grasping: (1) multi-level similarity matching that combines semantic (LLM-assisted), geometric (novel C-FPFH descriptor), and dimensional (SOBB) cues to retrieve candidate models from a small database of fewer than 100 YCB-derived models; (2) imitative grasp planning that transfers preplanned grasps from the best-matching candidate via a plane-detection-based point cloud registration; and (3) a two-stage stability-aware grasp fine-tuning step. The authors report real-world experiments on isolated and cluttered scenes, comparing with PointNetGPD, 3DSGrasp, a previous similarity approach, GraspNet, and HGGD, and claim significantly higher grasp success rates and robustness to sensing noise and environmental changes.

Significance. If the central claim were properly supported, this would be a valuable contribution: a learning-free alternative to large-scale grasp synthesis, with a new geometric descriptor (C-FPFH) for partial-to-complete point cloud similarity, a novel SOBB for dimensional matching, and a fine-tuning module that demonstrably improves grasp stability. The paper includes extensive real-world experiments, a useful ablation study, and an adversarial-object hold-out in Appendix B that provides some genuine evidence for the approach. However, the main evaluation is compromised: the matching experiments deliberately insert the test objects' original 3D models into the database (Section IV-B), and the grasping experiments on 'previously unseen objects' never demonstrate that those objects are absent from the database. Consequently, the headline numbers (96% isolated-object GSR, 93.5%/89.4% cluttered GSR) may reflect exact-model retrieval and pose registration rather than similarity-based generalization to unseen shapes. The paper also lacks statistical rigor (4-10 trials per cell, no confidence intervals or significance tests) and releases no code or data.

major comments (5)
  1. [§IV-B, Tables I and II] The similarity matching evaluation is a self-retrieval test, not a test of generalization to unknown objects. The text states that the authors 'utilize 5 different types of experimental objects with their original 3D models ... and incorporate them into the database', and defines Matching Accuracy as whether the original model of the target object appears in the top-x candidate list. Because the database contains the exact test model, a high MA only demonstrates that the pipeline can recognize a known model from a partial view and retrieve it; it says nothing about retrieving a similar but non-identical model for an object that is genuinely absent from the database. The reported 88% (non-occluded) and 72% (occluded) accuracies and the ablation conclusions about the contribution of C-FPFH therefore do not support the claim that the method generalizes to unseen objects. A proper held-out evaluation, in which the target object's model is not a database member, is required.
  2. [§IV-C, Table III] The '10 previously unseen objects' are not demonstrated to be absent from the database. The database is described as 'no more than 100 object models derived from the YCB dataset', and the test objects in Fig. 9 are YCB-style objects, yet the paper neither lists the database membership nor the identities of the test objects, and it never verifies non-overlap. If any of the 10 test objects has its exact mesh in the database, the pipeline can register the observed partial view to the exact model and transfer its preplanned grasps, which would make the 96% GSR a measure of instance recognition and pose registration rather than similarity-based generalization. This is the load-bearing condition for the abstract's claim that the method 'significantly outperforms existing benchmarks in grasping a wide variety of unknown objects'. The authors must either prove that all grasping test objects were not in the database or re-run the experiments with a clean hold-out set; the Appendix B adversarial-object experiment is a step in the right direction but covers only three objects.
  3. [§IV-C, Tables III-V] The trial counts are too small and no statistical analysis is provided, so the phrase 'significantly outperforms' is not supported. For example, in Table III each object/method combination has only 4-10 attempts; a 10/10 success rate has a 95% confidence interval of roughly [69%, 100%], and the differences between methods are not characterized by any confidence interval, hypothesis test, or variance measure. The aggregate GSR differences (96% vs 78% vs 67%) may reflect a real effect, but as reported they do not demonstrate statistical significance. The authors should report confidence intervals, per-object success counts with total attempts, and justify the number of trials, especially when claiming superiority over learning-based benchmarks.
  4. [§III-E and §III-F] The matching thresholds and free parameters are hand-set without sensitivity analysis. Section III-E assumes QS > 0.9 and DS < 0.1, Section III-F assumes SS < 0.1, and the C-FPFH pipeline uses fixed values for the voxel size (1.5 cm), FPFH neighborhood radius (1 cm), clustering radius (2 cm), normal-orientation difference (20°), and fine-tuning parameters k1 = 100, k2 = 5. Appendix A-C tests only two depth-filter settings on a single bowl and does not vary these parameters. Since candidate selection and, ultimately, grasp quality depend directly on these thresholds, the claimed robustness of the method across 'varying object types and operating environments' would be much better supported by a systematic sensitivity study or at least a discussion of how stable the results are to threshold changes.
  5. [§IV-B and §IV-E, Tables I and V] The ablation results do not consistently support the centrality of the C-FPFH descriptor. In non-occluded scenes, the baseline without geometric matching (w/o GM) achieves 90% GSR and 84% MA, essentially matching the full method (92% and 88%), and the text acknowledges that 'the database contains the categories of all experimental objects, making semantic and dimensional matching alone sufficient for achieving accurate results'. The C-FPFH descriptor's advantage is only clearly visible in occluded scenes. This is not internally contradictory, but it tempers the paper's claim that C-FPFH is the key enabling component for similarity matching. The authors should present per-condition results and clearly separate the contribution of each matching level, especially given the retrieval contamination noted above, so that readers can assess which components are truly needed for the headline performance.
minor comments (6)
  1. [§I Contribution list] There is a typo: 'Large Lanuage Models' should be 'Large Language Models'.
  2. [§IV-B] The manual marker placement for occluded scenes is described only in the caption of Fig. 8; it should be described in the main text, including how the marker is used to filter the segmentation output.
  3. [Table III] The APT values are reported as approximate single numbers (e.g., '≈6s', '≈2s') without variance or number of successful trials; consider reporting mean ± standard deviation over successful grasps.
  4. [§IV-E] The sentence 'The offset is disregarded in cases of grasp failure' should appear before the definition of AO, and the text should clarify whether the reported AO is averaged only over successful grasps or over all attempts.
  5. [Appendix C] The Word2Vec pre-filtering in the large-scale validation appears to contradict the criticisms of word-embedding semantic similarity in Section III-D; the authors should explain why a method rejected for semantic matching is acceptable as a pre-filter.
  6. [§V Conclusions] The conclusion states that the method 'significantly outperforms existing benchmarks across all metrics'; given the small sample sizes and lack of statistical tests, this wording should be softened to 'outperforms in the reported experiments'.

Circularity Check

1 steps flagged · score 6.0 of 10

Validation of the matching module is self-retrieval by construction; isolated-object generalization is not shown to be independent of database membership.

  1. self definitional [Section IV-B (Evaluation of Similarity Matching), with consequences for Section IV-C (Grasping Isolated Objects)]
    "we utilize 5 different types of experimental objects with their original 3D models, as shown in Fig. 8, and incorporate them into the database. During the matching process, if the original model of the target object is identified within the highest priority list of candidate models, we consider it an accurate match."

    The Matching Accuracy metric is defined as retrieval of the original model after that same original model has been deliberately inserted into the database; the reported 88%/72% accuracies are self-retrieval rates by construction, so the query and the target candidate are identical meshes. This does not validate the advertised non-identical-object similarity capability. The grasping section calls its 10 objects 'previously unseen' without checking absence from the <100-model YCB-derived database, and the cluttered and ablation trials reuse those objects; if their meshes are present, the top match is exact-model registration and the 96% GSR is identity lookup rather than similarity-based generalization.

full rationale

The paper is an empirical pipeline rather than a derivation, and most of the system components (LLM semantic scoring, C-FPFH histograms, SOBB, PDM registration, fine-tuning) are evaluated against external learning-based baselines in real robot trials; those comparisons are not circular. The central circularity is in the validation of the similarity-matching module: Section IV-B inserts the target objects' original 3D models into the database and then defines Matching Accuracy as whether the original model appears among the top candidates, making the reported accuracy a self-retrieval rate by construction. The grasping experiments in Sections IV-C through IV-E then refer to the test objects as 'previously unseen' without demonstrating that their YCB-derived meshes are absent from the YCB-derived database, and the cluttered/ablation trials reuse the same objects; if the exact meshes are present, the planner can copy preplanned grasps instead of generalizing across non-identical shapes. The comparison with the authors' earlier similarity method [14] is a self-citation but is used as a fair baseline rather than as load-bearing justification. Appendix B's adversarial objects (headphone, adapter, gimbal) provide a genuine hold-out with 90% success vs. 40% for GraspNet, which prevents the circularity from being total. The score of 6 reflects that one key reported quantity (matching accuracy, and potentially the isolated-object GSR) reduces in part to identity retrieval rather than to similarity-based prediction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 3 invented entities

The central claim rests on the assumption that every target has a good database analogue, and the paper's own evaluation makes that assumption trivially true by adding the exact test models. The new algorithmic constructs (C-FPFH, SOBB, PDM) are internally plausible but are only validated on the paper's small object set with many hand-set thresholds.

free parameters (7)
  • QS threshold = >0.9
    Geometric matching selects candidates only when QS exceeds this empirical cutoff (Section III-E).
  • DS threshold = <0.1
    Distributional similarity cutoff for geometric matching (Section III-E).
  • SS threshold = <0.1
    Dimensional matching cutoff for bounding-extent differences (Section III-F).
  • C-FPFH sampling and neighborhood radii = voxel 1.5 cm, FPFH radius 1 cm, cluster radius 2 cm
    Hand-set scales controlling local feature extraction and clustering (Section III-E).
  • Fine-tuning parameters = k1=100, k2=5, angle thresholds 20, 40, 10 degrees
    Empirical values for grasp classification and position adjustment (Section III-J).
  • Large-scale prefiltering parameters = delta=50%, alpha=0.5, beta=5, voxel 5 mm
    Hand-set ratios and voxel size for subset selection in the Dex-Net validation (Appendix B-C).
  • Normal-orientation difference in clustering = 20 degrees
    Empirical tolerance for cluster growth in C-FPFH descriptor construction (Section III-E).
assumptions (5)
  • domain assumption The database contains a model similar enough to every target object.
    The method transfers preplanned grasps from matched models; without such coverage, Section III-G falls back to non-similar models and planning may fail. Section IV-B strengthens this by inserting exact test models into the database.
  • domain assumption Single-view point cloud is a usable partial representation of the target.
    The whole pipeline relies on background subtraction, instance segmentation, and pixel-to-point projection producing a point cloud that overlaps with the visible surfaces of the target (Section III-B).
  • domain assumption Table-top plane normal is known or estimable.
    SOBB generation fixes one axis to the support plane normal; the horizontal table assumption is stated, and non-flat surfaces require the support surface to be fully visible (Section III-F, Appendix B-A).
  • ad hoc to paper Two dominant FPFH components are stable under sensing noise.
    This is the central hypothesis behind C-FPFH and is validated only empirically in Appendix A/B, not proven for complex surfaces; the paper itself notes the gap shrinks for intricate geometries.
  • standard math Background algorithms work as specified.
    FPFH, PCA, ICP, RANSAC, ball-pivoting, and plane detection are treated as reliable off-the-shelf tools (Sections III-E, III-H, III-I).
invented entities (3)
  • C-FPFH descriptor
    purpose: Aggregates dominant FPFH component pairs and their spatial clusters into QS/DS metrics for partial-to-complete point-cloud similarity.
    Validated only on the paper's own objects and thresholds; no external benchmark or formal characterization is provided.
  • Semi-Oriented Bounding Box (SOBB)
    purpose: Estimates target dimensions from partial point clouds by fixing one axis to the support-plane normal.
    Defined in this paper and evaluated only through the dimensional matching experiments; no external validation.
  • PDM-based point cloud registration
    purpose: Uses detected planes to seed coarse alignment of partial and complete clouds before ICP.
    The paper's own comparison (Fig. 6d) shows cases where RANSAC+ICP works better; no independent evaluation is offered.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Multi-Level Similarity Approach for Single-View Object Grasping: Matching, Planning, and Fine-Tuning." pith.science (2026). https://pith.science/paper/BPVKFJFW

@misc{pith2026250711938,
  author       = {Pith},
  title        = {Pith review of: A Multi-Level Similarity Approach for Single-View Object Grasping: Matching, Planning, and Fine-Tuning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BPVKFJFW}},
  note         = {Machine review of arXiv:2507.11938}
}
read the original abstract

Grasping unknown objects from a single view has remained a challenging topic in robotics due to the uncertainty of partial observation. Recent advances in large-scale models have led to benchmark solutions such as GraspNet-1Billion. However, such learning-based approaches still face a critical limitation in performance robustness for their sensitivity to sensing noise and environmental changes. To address this bottleneck in achieving highly generalized grasping, we abandon the traditional learning framework and introduce a new perspective: similarity matching, where similar known objects are utilized to guide the grasping of unknown target objects. We newly propose a method that robustly achieves unknown-object grasping from a single viewpoint through three key steps: 1) Leverage the visual features of the observed object to perform similarity matching with an existing database containing various object models, identifying potential candidates with high similarity; 2) Use the candidate models with pre-existing grasping knowledge to plan imitative grasps for the unknown target object; 3) Optimize the grasp quality through a local fine-tuning process. To address the uncertainty caused by partial and noisy observation, we propose a multi-level similarity matching framework that integrates semantic, geometric, and dimensional features for comprehensive evaluation. Especially, we introduce a novel point cloud geometric descriptor, the C-FPFH descriptor, which facilitates accurate similarity assessment between partial point clouds of observed objects and complete point clouds of database models. In addition, we incorporate the use of large language models, introduce the semi-oriented bounding box, and develop a novel point cloud registration approach based on plane detection to enhance matching accuracy under single-view conditions. Videos are available at https://youtu.be/qQDIELMhQmk.

Figures

Figures reproduced from arXiv: 2507.11938 by the authors.

Figure 1
Figure 1. An example of grasping a novel object using our three-step method. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our proposed approach. The system inputs (shown in blue boxes) include a single-view RGBD image and an existing object model [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Three-step generation of C-FPFH descriptors for object point clouds. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (19 more)
Figure 4
Figure 4. Figure 4: Test results of using C-FPFH-based geometric matching to identify similar models from single-view object point clouds. All candidate models are [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Comparison of different types of bounding boxes applied to a single [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: (a)-(c): Three steps to perform PDM-based point cloud registration [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Two-stage grasp fine-tuning, including grasp position and grasp center [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: The experimental objects used for similarity matching. The first and [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Grasping various novel objects using our similarity-based method. The top left corner of each figure displays the single-view object point cloud, the [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Comparison of different grasp planning methods based on object point clouds. The blue grasps represent all the generated grasp candidates, while [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]
Figure 11
Figure 11. Figure 11: Objects in clutter used for grasping experiments. Each clutter set [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 12
Figure 12. Figure 12: Comparison of our method with grasping benchmarks for objects in clutter. In the outputs of GraspNet and HGGD, red grasps indicate high-quality [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]
Figure 13
Figure 13. Figure 13: Evaluation of grasp quality using a graph paper. After a pick-and [PITH_FULL_IMAGE:figures/full_fig_p013_13.png]
Figure 14
Figure 14. Figure 14: Application of our method to achieve grasping tasks in three different scenarios without the original models of the environmental objects. Handover 8 10 14 [PITH_FULL_IMAGE:figures/full_fig_p014_14.png]
Figure 15
Figure 15. Figure 15: Comparison of the performance robustness of different methods. [PITH_FULL_IMAGE:figures/full_fig_p014_15.png]
Figure 16
Figure 16. Figure 16: Failures of learning-based methods in on-shelf and hand-over tasks. GraspNet HGGD [PITH_FULL_IMAGE:figures/full_fig_p014_16.png]
Figure 17
Figure 17. Figure 17: Matching cases where two objects are similar in two out of three [PITH_FULL_IMAGE:figures/full_fig_p015_17.png]
Figure 18
Figure 18. Figure 18: Using different models from the top candidates with high fitness [PITH_FULL_IMAGE:figures/full_fig_p015_18.png]
Figure 19
Figure 19. Figure 19: Feature extraction of sampled points from different object regions, [PITH_FULL_IMAGE:figures/full_fig_p016_19.png]
Figure 20
Figure 20. Figure 20: Validations of the robustness of C-FPFH-based geometric matching [PITH_FULL_IMAGE:figures/full_fig_p016_20.png]
Figure 21
Figure 21. Figure 21: SOBB generation and grasp planning for objects on non-flat surfaces. GraspNe [PITH_FULL_IMAGE:figures/full_fig_p017_21.png]
Figure 23
Figure 23. Figure 23: Object point clouds after being downsampled with the same voxel [PITH_FULL_IMAGE:figures/full_fig_p018_23.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 41 canonical work pages

  1. [1]

    Deep learning approaches to grasp synthesis: A review,

    R. Newbury, M. Gu, L. Chumbley, A. Mousavian, C. Eppner, J. Leitner, J. Bohg, A. Morales, T. Asfour, D. Kragic, D. Fox, and A. Cosgun, “Deep learning approaches to grasp synthesis: A review,” IEEE Trans- actions on Robotics , vol. 39, no. 5, pp. 3994–4015, 2023

    work page 2023

  2. [2]

    Learning ambidextrous robot grasping policies,

    J. Mahler, M. Matl, V . Satish, M. Danielczuk, B. DeRose, S. McKinley, and K. Goldberg, “Learning ambidextrous robot grasping policies,” Science Robotics, vol. 4, no. 26, p. eaau4984, 2019

    2019

  3. [3]

    Closing the loop for robotic grasping: A real-time, generative grasp synthesis approach,

    D. Morrison, P. Corke, and J. Leitner, “Closing the loop for robotic grasping: A real-time, generative grasp synthesis approach,” in RSS, 2018

    work page 2018

  4. [4]

    Learning 6-dof grasping interaction via deep geometry- aware 3d representations,

    X. Yan, J. Hsu, M. Khansari, Y . Bai, A. Pathak, A. Gupta, J. Davidson, and H. Lee, “Learning 6-dof grasping interaction via deep geometry- aware 3d representations,” in ICRA, 2018, pp. 3766–3773

    work page 2018

  5. [5]

    Robotic pick-and-place of novel objects in clutter with multi-affordance grasping and cross-domain image matching,

    A. Zeng, S. Song, K.-T. Yu, E. Donlon, F. R. Hogan, M. Bauza, D. Ma, O. Taylor, M. Liu, E. Romo, N. Fazeli, F. Alet, N. C. Dafle, R. Holladay, I. Morena, P. Qu Nair, D. Green, I. Taylor, W. Liu, T. Funkhouser, and A. Rodriguez, “Robotic pick-and-place of novel objects in clutter with multi-affordance grasping and cross-domain image matching,” The Internat...

    work page 2019

  6. [6]

    Grasp pose detection in point clouds,

    A. ten Pas, M. Gualtieri, K. Saenko, and R. Platt, “Grasp pose detection in point clouds,” The International Journal of Robotics Research, vol. 36, no. 13-14, pp. 1455–1473, 2017

    work page 2017

  7. [7]

    6-dof graspnet: Variational grasp generation for object manipulation,

    A. Mousavian, C. Eppner, and D. Fox, “6-dof graspnet: Variational grasp generation for object manipulation,” in ICCV, 2019

    work page 2019

  8. [8]

    V olumetric grasping network: Real-time 6 dof grasp detection in clutter,

    M. Breyer, J. J. Chung, L. Ott, S. Roland, and N. Juan, “V olumetric grasping network: Real-time 6 dof grasp detection in clutter,” in CoRL, 2020

    work page 2020

Show all 43 references
  1. [9]

    Synergies between affordance and geometry: 6-dof grasp detection via implicit representa- tions,

    Z. Jiang, Y . Zhu, M. Svetlik, K. Fang, and Y . Zhu, “Synergies between affordance and geometry: 6-dof grasp detection via implicit representa- tions,” in RSS, 2021

    2021

  2. [10]

    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 CVPR, 2020, pp. 11 444–11 453

    2020

  3. [11]

    Grasp-anything: Large-scale grasp dataset from foundation models,

    A. D. Vuong, M. N. Vu, H. Le, B. Huang, H. T. T. Binh, T. V o, A. Kugi, and A. Nguyen, “Grasp-anything: Large-scale grasp dataset from foundation models,” in ICRA, 2024, pp. 14 030–14 037

    2024

  4. [12]

    The columbia grasp database,

    C. Goldfeder, M. Ciocarlie, H. Dang, and P. K. Allen, “The columbia grasp database,” in ICRA, 2009, pp. 1710–1716

    2009

  5. [13]

    Ontology-assisted generalisation of robot action execution knowledge,

    A. Mitrevski, P. G. Pl ¨oger, and G. Lakemeyer, “Ontology-assisted generalisation of robot action execution knowledge,” in IROS, 2021, pp. 6763–6770. 19

    2021

  6. [14]

    Category-association based similarity matching for novel object pick-and-place task,

    H. Chen, T. Kiyokawa, W. Wan, and K. Harada, “Category-association based similarity matching for novel object pick-and-place task,” IEEE Robotics and Automation Letters , vol. 7, no. 2, pp. 2961–2968, 2022

    2022

  7. [15]

    Pointnetgpd: Detecting grasp configurations from point sets,

    H. Liang, X. Ma, S. Li, M. G ¨orner, S. Tang, B. Fang, F. Sun, and J. Zhang, “Pointnetgpd: Detecting grasp configurations from point sets,” in ICRA, 2019, pp. 3629–3635

    2019

  8. [16]

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

    R. Q. Charles, H. Su, M. Kaichun, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” in CVPR, 2017, pp. 77–85

    2017

  9. [17]

    Qt-opt: Scalable deep reinforcement learning for vision-based robotic manipulation,

    D. Kalashnikov, A. Irpan, P. Pastor, J. Ibarz, A. Herzog, E. Jang, D. Quillen, E. Holly, M. Kalakrishnan, V . Vanhoucke, and S. Levine, “Qt-opt: Scalable deep reinforcement learning for vision-based robotic manipulation,” in CoRL, vol. 87, 2018, pp. 651–673

    2018

  10. [18]

    3dsgrasp: 3d shape-completion for robotic grasp,

    S. S. Mohammadi, N. F. Duarte, D. Dimou, Y . Wang, M. Taiana, P. Morerio, A. Dehban, P. Moreno, A. Bernardino, A. Del Bue, and J. Santos-Victor, “3dsgrasp: 3d shape-completion for robotic grasp,” in ICRA, 2023, pp. 3815–3822

    2023

  11. [19]

    Scarp: 3d shape completion in arbitrary poses for improved grasping,

    B. Sen, A. Agarwal, G. Singh, B. B., S. Sridhar, and M. Krishna, “Scarp: 3d shape completion in arbitrary poses for improved grasping,” in ICRA, 2023, pp. 3838–3845

    2023

  12. [20]

    Efficient heatmap- guided 6-dof grasp detection in cluttered scenes,

    S. Chen, W. Tang, P. Xie, W. Yang, and G. Wang, “Efficient heatmap- guided 6-dof grasp detection in cluttered scenes,” IEEE Robotics and Automation Letters, vol. 8, no. 8, pp. 4895–4902, 2023

    2023

  13. [21]

    Model-free and learning-free grasping by local contact moment matching,

    M. Adjigble, N. Marturi, V . Ortenzi, V . Rajasekaran, P. Corke, and R. Stolkin, “Model-free and learning-free grasping by local contact moment matching,” in IROS, 2018, pp. 2933–2940

    2018

  14. [22]

    Robust feature detection and local classification for surfaces based on moment analysis,

    U. Clarenz, M. Rumpf, and A. Telea, “Robust feature detection and local classification for surfaces based on moment analysis,” IEEE Transactions on Visualization and Computer Graphics , vol. 10, no. 5, pp. 516–524, 2004

    2004

  15. [23]

    Learning-free grasping of unknown objects using hidden superquadrics,

    Y . Wu, W. Liu, Z. Liu, and G. S. Chirikjian, “Learning-free grasping of unknown objects using hidden superquadrics,” in RSS, 2023

    2023

  16. [24]

    Robust and accurate superquadric recovery: a probabilistic approach,

    W. Liu, Y . Wu, S. Ruan, and G. S. Chirikjian, “Robust and accurate superquadric recovery: a probabilistic approach,” in CVPR, 2022, pp. 2666–2675

    2022

  17. [25]

    Robust grasp detection with incomplete point cloud and complex background,

    X. Wang, S. Nisar, and F. Matsuno, “Robust grasp detection with incomplete point cloud and complex background,” Advanced Robotics, vol. 35, no. 10, pp. 619–634, 2021

    2021

  18. [26]

    Template-based learning of grasp selection,

    A. Herzog, P. Pastor, M. Kalakrishnan, L. Righetti, T. Asfour, and S. Schaal, “Template-based learning of grasp selection,” in ICRA, 2012, pp. 2379–2384

    2012

  19. [27]

    Detecting twenty-thousand classes using image-level supervision,

    X. Zhou, R. Girdhar, A. Joulin, P. Kr ¨ahenb¨uhl, and I. Misra, “Detecting twenty-thousand classes using image-level supervision,” in ECCV, 2022, pp. 350–368

    2022

  20. [28]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, P. Doll´ar, and R. Girshick, “Segment anything,” in ICCV, 2023, pp. 4015–4026

    2023

  21. [29]

    Efficient estimation of word representations in vector space,

    T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient estimation of word representations in vector space,” in ICLR, 2013, p. 112

    2013

  22. [30]

    Glove: Global vectors for word representation,

    J. Pennington, R. Socher, and C. Manning, “Glove: Global vectors for word representation,” in EMNLP, vol. 14, 2014, pp. 1532–1543

    2014

  23. [31]

    Gpt-4: Openai’s generative pre-trained transformer, version 4,

    OpenAI, “Gpt-4: Openai’s generative pre-trained transformer, version 4,” 2024. [Online]. Available: https://openai.com/research/gpt-4

    2024

  24. [32]

    Benchmarking in manipulation research: Using the yale-cmu- berkeley object and model set,

    B. Calli, A. Walsman, A. Singh, S. Srinivasa, P. Abbeel, and A. M. Dollar, “Benchmarking in manipulation research: Using the yale-cmu- berkeley object and model set,” IEEE Robotics & Automation Magazine, vol. 22, no. 3, pp. 36–52, 2015

    2015

  25. [33]

    Fast point feature histograms (fpfh) for 3d registration,

    R. B. Rusu, N. Blodow, and M. Beetz, “Fast point feature histograms (fpfh) for 3d registration,” in ICRA, 2009, pp. 3212–3217

    2009

  26. [34]

    Cad-model recognition and 6dof pose estimation using 3d cues,

    A. Aldoma, M. Vincze, N. Blodow, D. Gossow, S. Gedikli, R. B. Rusu, and G. Bradski, “Cad-model recognition and 6dof pose estimation using 3d cues,” in ICCV, 2011, pp. 585–592

    2011

  27. [35]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,

    M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography,” Communications of the ACM , vol. 24, pp. 381–395, 1981

    1981

  28. [36]

    Object modeling by registration of multiple range images,

    Y . Chen and G. Medioni, “Object modeling by registration of multiple range images,” in ICRA, vol. 3, 1991, pp. 2724–2729

    1991

  29. [37]

    A robust statistics approach for plane detection in unorganized point clouds,

    A. M. C. Ara ´ujo and M. M. Oliveira, “A robust statistics approach for plane detection in unorganized point clouds,” Pattern Recognition, vol. 100, p. 107115, 2020

    2020

  30. [38]

    Planning grasps with suction cups and parallel grippers using superimposed segmentation of object meshes,

    W. Wan, K. Harada, and F. Kanehiro, “Planning grasps with suction cups and parallel grippers using superimposed segmentation of object meshes,” IEEE Transactions on Robotics , vol. 37, pp. 166–184, 2021

    2021

  31. [39]

    The wrs robot planning and control system,

    W. Wan and et al., “The wrs robot planning and control system,” 2024. [Online]. Available: https://github.com/wanweiwei07/wrs

    2024

  32. [40]

    Ikfast: A fast and robust inverse kinematics solver for robotics,

    R. Diankov and P. Kormushev, “Ikfast: A fast and robust inverse kinematics solver for robotics,” in IROS, 2010, pp. 1–8

    2010

  33. [41]

    The ball-pivoting algorithm for surface reconstruction,

    F. Bernardini, J. Mittleman, H. Rushmeier, C. Silva, and G. Taubin, “The ball-pivoting algorithm for surface reconstruction,” IEEE Transactions on Visualization and Computer Graphics , vol. 5, no. 4, pp. 349–359, 1999

    1999

  34. [42]

    Hao Chen (Member, IEEE) received the B.Eng

    Panda3D Community, Panda3D Manual, Carnegie Mellon University, 2024, https://docs.panda3d.org/. Hao Chen (Member, IEEE) received the B.Eng. degree from Zhejiang University, China, in 2019, and the M.S. degree from the Graduate School of Engineering Science, Osaka University, J...

    2024

  35. [1997]

    From 2005 to 2006, he was a Visiting Scholar at the Computer Science Department, Stanford University, California, USA

    From 1997 to 2002, he was a Research As- sociate at Hiroshima University, Hiroshima, Japan. From 2005 to 2006, he was a Visiting Scholar at the Computer Science Department, Stanford University, California, USA. He was a Research scientist at Na- tional Institute of Advanced In...

    1997

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.