3D-CDRGP: Towards Cross-Device Robotic Grasping Policy in 3D Open World

Chengrui Zhang; Chenru Jiang; Jie Sun; Kaizhu Huang; Rui Zhang; Weiguang Zhao; Yuyao Yan

REVIEW 3 major objections 6 minor 1 cited by

3D-CDRGP: Towards Cross-Device Robotic Grasping Policy in 3D Open World

T0 review · 3 major / 6 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 A frozen 3D detector, with no retraining, can grasp unseen objects on new camera-and-robot hardware, the paper claims, and two real cross-manufacturer systems back it up.

desk verdict A genuine training-free cross-device grasping demo built on a softmax-background heuristic that is plausible but under-validated; deserves referee time but not yet a trusted open-world claim. read the letter →

arxiv 2411.18133 v2 pith:7OQTSSAQ submitted 2024-11-27 cs.RO cs.CV

Weiguang Zhao , Chenru Jiang , Chengrui Zhang , Jie Sun , Yuyao Yan , Rui Zhang , Kaizhu Huang This is my paper · ORCID

classification cs.ROcs.CV

keywords cross-deviceroboticgrasping3Dopenworldcategory-agnosticobjectdetectionpointcloudclusteringclosed-setSoftMaxforeground-backgroundsplitwithoutretraining

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 is trying to establish that a 3D object detector trained on a closed set of categories can be transplanted onto entirely different camera-and-robot systems and used to grasp objects it has never seen, with no retraining or fine-tuning. If true, a single pretrained detector can serve as a plug-and-play perception module for robotic picking across hardware generations, manufacturers, and object types. The policy replaces the detector's multi-class semantic output with a binary foreground/background split (using the observed tendency of the background class to dominate the SoftMax), clusters foreground points by geometry, and scores proposals with a revised rule that includes point count and height. The authors report near-perfect recognition and grasp rates on two real systems built from different vendors' cameras and arms, including stacked and everyday objects absent from the training set.

What carries the argument

SSGC-Seg is the load-bearing module: it turns the multi-class semantic SoftMax output of a frozen closed-set detector into a binary foreground/background decision via Eqs. (1)–(2), exploiting the observed dominance of the background class, and then clusters the foreground points with the parameter-free geo-clustering of PBNet in the original coordinate system to form category-agnostic object proposals. ScoreNet‡ revises each proposal's confidence by combining the learned backbone feature score with the proposal's point count and mean height (Eq. 4), discarding proposals below a point-count threshold and guiding the gripper toward higher objects. Hand-eye calibration (Eq. 5) and a yaw-angle rule (Eq. 6) convert the proposals into robot grasps.

What would settle it

Run the frozen detector through SSGC-Seg on a scene with an unseen object that is similar in shape to a training class but placed on a cluttered or textured background (for instance, a doll resting on a patterned cloth), and measure the fraction of foreground points for which the original multi-class SoftMax assigns the top value to the background class; if that fraction is not near one, the proposal is fragmented and the reported near-100% grasp rates would not transfer to that scene.

Watch

Extended reading notes

Core claim

The central claim is that the gap between closed-set 3D detection and open-world cross-device grasping can be closed without learning. The paper shows that a pretrained semantic branch, trained on a closed set of object classes, still behaves as a reliable background detector because the background class consistently receives the highest SoftMax value. Converting the multi-class SoftMax into a binary foreground/background score — the background score versus the maximum over all non-background classes — and clustering the resulting foreground points by geometry yields object proposals for arbitrary object categories. A revised score function, ScoreNet‡, adds point count and mean height to the learned confidence score, filtering out fragmentary false detections and prioritizing taller objects for top-down grasping. Across two complete robot–camera systems from different manufacturers (an AiNSTEC PRO camera with an Elite EC66 arm, and a RealSense D455 with a Dobo CR robot), the policy grasps previously unseen and stacked objects without retraining.

Load-bearing premise

The load-bearing premise is that the pretrained semantic branch always gives the background class the highest SoftMax value, so that the binary-by-SoftMax split of Eqs. (1)–(2) cleanly separates graspable foreground points from the tabletop; if an unseen object makes the network spread probability across known classes or weakens the background's dominance, the proposals fragment and the grasp sequence misses targets.

Editorial extensions

If this is right

A pretrained closed-set 3D detector can be deployed on new camera hardware — different brands, resolutions, depth accuracies — with no data collection or fine-tuning.
The grasping system extends to arbitrary unseen object categories, not just training classes, as long as objects are separable from the background by geometry.
The policy is compatible with most clustering-based 3D instance-segmentation methods, so a stronger detector can be swapped in without changing the grasping pipeline.
Deployment cost drops: new robot arms and cameras only require kinematic setup and hand-eye calibration, while the perception model stays frozen.
The revised scoring rule (point count plus height) also suppresses fragmentary false detections and improves ordering in stacked scenes.

Reading between the lines

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

The background-dominance observation is a property of the specific pretrained model and its training distribution; the paper does not test whether the same dominance holds for models trained on other closed sets, so the policy's open-world generality across arbitrary detectors remains an open question.
A testable extension is to replace the hard argmax split with an uncertainty-aware criterion such as SoftMax entropy, which might preserve the gain when the background score is not strictly maximal.
Because deployment is training-free, the pipeline could be paired with vision-language models or open-vocabulary detectors to add object selection, turning open-category grasping into open-instruction grasping; the authors mention this as future work.
Clustering in original coordinates avoids offset-branch errors under domain shift but also discards learned object-center priors; on heavily occluded or interlocked scenes, that trade-off could reverse.

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.

Desk Editor's Note

read the letter

Dear colleague,

This paper is worth a serious look if you work on robotic grasping or open-world 3D detection. The core idea is simple: take a closed-set 3D instance segmentation model, treat “background” as the class whose softmax score beats every seen object class, and then cluster the remaining points geometrically to get graspable object proposals. They add a modified scoring function that prefers larger, higher objects, and they show the whole pipeline running on two different camera/robot combinations without retraining. The real-robot demos are genuine and the reported recognition/grasp rates are high. The cross-device angle is genuinely new — most prior work trains on augmented point clouds or adapts per hardware, and a training-free transfer is a useful thing to demonstrate.

The paper does several things well. The SSGC-Seg transformation (Eq. 1–2) is a neat adaptation of a closed-set detector to category-agnostic detection, and the authors are honest that it rests on an empirical observation about softmax behavior rather than a proof. The comparison of two clustering methods on the Boxes dataset is fine, and PBNet’s binary clustering being better than PointGroup’s distance clustering is a plausible finding. The writing is clear, and the supplementary videos (at least as described) back up the numbers.

The soft spots are real. First, the central assumption — that the background class always dominates for background points and that unseen foreground objects will trigger a seen-class softmax above background — is not stress-tested. The closed-set Boxes benchmark uses the same rectangular block categories that the model was pretrained on, so it cannot validate open-world behavior. The real-world scenarios use only a handful of objects, with no trial counts, no per-scene variance, and no failure analysis beyond one smooth spray can slipping. Second, there are four hand-set thresholds (alpha, N_theta, C_theta, d_theta) whose sensitivity is never examined; the high success rates are therefore partly fitted to the setups. Third, there are no baselines on the real systems — no closed-set detector without SSGC-Seg, no existing open-set detector, no retrained alternative. That makes it hard to know how much of the success comes from the specific SSGC-Seg trick versus the geometric clustering and the top-down grasping prior.

None of this makes the paper a reject in my view. The demonstration is real, the method is coherent, and the limitation section acknowledges the main gaps. But the open-world generalization claim is not yet fully supported. The authors should release code and data, report trial counts and variance, run at least one baseline on the same hardware, and analyze threshold sensitivity and the softmax-dominance condition on truly unseen categories.

I’d send it to review, with the expectation of major revision. For you as a reader: worth a look for the idea, but do not treat the 100% numbers as established.

Referee Report

3 major / 6 minor

Summary. The paper proposes 3D-CDRGP, a training-free policy that adapts closed-set clustering-based 3D object detectors to cross-device robotic grasping in open-world scenes. The method introduces SSGC-Seg, which converts point-wise semantic softmax scores into a binary foreground/background prediction using Eqs. (1)-(2), then clusters foreground points in the original coordinate system and scores proposals with ScoreNet-dagger, a modified scoring function that combines the backbone confidence with point count and mean height. The authors evaluate the approach on the Boxes dataset using PointGroup and PBNet as base detectors, and on two real robot systems (Elite EC66/AiNSTEC PRO and Dobo CR/RealSense D455) across five scenarios, reporting high recognition and grasping rates. The paper claims plug-and-play open-category grasping across devices without retraining or fine-tuning.

Significance. If the central claims hold, the paper offers a practical and deceptively simple solution to a real deployment problem: reusing a frozen closed-set 3D detector on new camera/robot hardware for category-agnostic grasping. The approach is compatible with multiple clustering-based detectors, and the authors demonstrate it on two genuinely different hardware systems with realistic objects, which is more than many grasping papers do. The decision to rely on geometry (point counts, height) rather than learned features for scoring is a sensible heuristic that is easy to reproduce. However, the evidence as presented is not yet sufficient: the key softmax-dominance premise behind SSGC-Seg is unvalidated on truly unseen categories, the real-world evaluation has very small implicit trial counts and no baselines on the same hardware, and the four manually chosen parameters are not ablated. The paper's value will be much stronger if these gaps are closed; as it stands, the central claim is plausible but under-supported.

major comments (3)

[SSGC-Seg, Eqs. (1)-(2)] The open-world claim rests on the assumption that for every foreground point on an unseen object, max(Si[2:M]) > Si[1], and for background points the reverse. The paper only states 'we observe' this behavior; it provides no quantitative evidence, no per-point foreground/background accuracy, and no analysis of how the softmax margin behaves as a function of object category, sensor, or distance. On the Boxes benchmark (Tab. 1) all objects are from the closed-set pre-training categories, so the rule's good performance there does not validate open-world operation. Since a violation of this assumption assigns an entire object to background (no proposal, no grasp) or table points to foreground (false proposals), the central 'open-category, cross-device' claim is currently unsupported. Please add a direct validation on held-out unseen categories, including per-point segmentation metrics and failure cases, or otherwise bound the condition.
[Table 2 and 'Cross-Device Robotic Grasping Comparison'] The real-world evaluation reports only aggregate recognition and grasping rates, with no trial counts, no repeated runs, and no per-object breakdown. The single failure in Scenario 4 (spray can slipping) implies a denominator of nine trials (8/9 = 88.9%), which is too small to support the 100% success claims elsewhere. Moreover, there is no comparison on the same hardware against an alternative policy (e.g., the original ScoreNet, a closed-set baseline, or a generic instance-segmentation grasping method), so it is unclear whether the proposed SSGC-Seg and ScoreNet-dagger are responsible for the results. Please report trial counts, standard deviations or at least raw per-trial outcomes, and a same-hardware baseline comparison.
[Eq. (4) and Algorithm 1] The pipeline has four manually set thresholds/weights (alpha=0.3, N_theta=60, C_theta=0.5, d_theta=2 in the Experiments section), and the reported success rates are conditional on these values. There is no ablation or sensitivity analysis showing how recognition/grasping rates vary with these parameters, nor any argument that the chosen values are robust across the two systems. If the method is intended as a plug-and-play policy, the parameter robustness should be demonstrated. Please add an ablation study, or at least sensitivity curves over plausible ranges for each parameter.

minor comments (6)

[Figure 2] The label 'SSGI-Seg' in the figure appears to be a typo for 'SSGC-Seg'; also, the arrow labels Fb, N_I^map, Hm are not all defined in the figure caption.
[Our Policy, Overview] The phrase 'the pre-training model of he closed-set' should read 'the pre-training model of the closed-set'.
[Eq. (6)] The notation yxmax and xymax is not defined clearly; the text should spell out that yxmax is the y-coordinate of the point with maximum x, and xymax is the x-coordinate of the point with maximum y.
[Algorithm 1] The threshold C_theta is used in line 13 but is not defined in the 'ScoreNet-dagger' section; please define it where the other parameters are introduced.
[Table 1] The Boxes rows report only the proposed 'PointGroup + Ours' and 'PBNet + Ours' without baseline results on the same dataset; please include the unmodified closed-set detectors for reference.
[SSGC-Seg section] The phrase 'data collection from 2th to Mth index' is awkward; consider rewriting as 'the slice of S from index 2 to M'.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: SSGC-Seg and ScoreNet‡ are heuristic adaptations, not derived predictions.

full rationale

The paper's derivation chain is not circular. SSGC-Seg (Eqs. 1–2) transforms the closed-set softmax outputs into a binary foreground/background score by comparing the background class probability with the maximum of all other classes. This is an explicit heuristic based on an empirical observation about the semantic branch's behavior; it is not a result derived from the target claim of open-world grasping. ScoreNet‡ (Eq. 4) combines the existing ScoreNet output with point count and mean height using manually specified thresholds (α=0.3, Nθ=60, Cθ=0.5, dθ=2). These are fixed configuration parameters, not fitted to the evaluation data in a way that makes the reported success rates equivalent to those parameter choices. The use of PBNet (Zhao et al. 2023) is a self-citation, but it is justified by a head-to-head comparison with PointGroup and other methods on ScanNetv2 (Table 1), providing independent evidence for the choice. The claimed open-world capability rests on an unproven but non-circular assumption about softmax behavior, and the real-world experiments are empirical demonstrations rather than predictions derived from the method itself. No step in the derivation reduces to its own inputs by construction, so circularity is minimal.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the background-softmax dominance assumption, the sufficiency of geometric clustering, and the accuracy of hand-eye calibration. The free parameters are the hand-set thresholds in the scoring and grasping pipeline, which are effectively fitted to the demonstrated scenarios.

free parameters (4)

alpha (α) = 0.3
Weight in ScoreNet‡ combining the original ScoreNet confidence and the mean height of the object proposal (Eq. 4). Chosen by hand, with no sensitivity analysis.
N_theta (Nθ) = 60
Point-count threshold below which a proposal is discarded as a false positive (Eq. 4). Set by hand and likely tuned to the test setups.
C_theta (Cθ) = 0.5
Minimum confidence score required for a grasp candidate in Algorithm 1. A hand-set threshold that affects which objects are grasped.
d_theta (dθ) = 2
Point density threshold used in Algorithm 1 to separate high-density and low-density point sets for clustering. Set by hand.

assumptions (3)

domain assumption The pre-trained semantic branch's background class has the highest SoftMax value for all points in open-world scenes
This is the basis of SSGC-Seg, Eq. (1)-(2). It is asserted as an observation but not proven or tested across diverse scenes, and is the weakest load-bearing premise.
domain assumption Geometry-based clustering in the original coordinate system is sufficient to separate object instances for grasping
The policy deliberately ignores predicted point offsets and clusters foreground points directly on original coordinates (Algorithm 1). This assumes that object instances are separable by density and geometry alone.
domain assumption Hand-eye calibration is accurate for the tested setups
Grasp center transformation (Eq. 5) depends on the rotation matrix R and translation vector T from calibration. If calibration is inaccurate, the robot misses the object.

how reviews work

0 comments

Cite this review

Pith. "Pith review of 3D-CDRGP: Towards Cross-Device Robotic Grasping Policy in 3D Open World." pith.science (2026). https://pith.science/paper/7OQTSSAQ

@misc{pith2026241118133,
  author       = {Pith},
  title        = {Pith review of: 3D-CDRGP: Towards Cross-Device Robotic Grasping Policy in 3D Open World},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7OQTSSAQ}},
  note         = {Machine review of arXiv:2411.18133}
}

read the original abstract

Given the diversity of devices and the product upgrades, cross-device research has become an urgent issue that needs to be tackled. To this end, we pioneer in probing the cross-device (cameras & robotics) grasping policy in the 3D open world. Specifically, we construct two real-world grasping setups, employing robotic arms and cameras from completely different manufacturers. To minimize domain differences in point clouds from diverse cameras, we adopt clustering methods to generate 3D object proposals. However, existing clustering methods are limited to closed-set scenarios, which confines the robotic graspable object categories and ossifies the deployment scenarios. To extend these methods to open-world settings, we introduce the SSGC-Seg module that enables category-agnostic 3D object detection. The proposed module transforms the original multi-class semantic information into binary semantic cues-foreground and background by analyzing the SoftMax value of each point, and then clusters the foreground points based on geometric information to form initial object proposals. Furthermore, ScoreNet{\ddag} is designed to score each detection result, and the robotic arm prioritizes grasping the object with the highest confidence score. Experiments on two different types of setups highlight the effectiveness and robustness of our policy for cross-device robotics grasping research. Our code is provided in the supplementary and will be released upon acceptance.

Figures

Figures reproduced from arXiv: 2411.18133 by the authors.

**Figure 2.** Overview of Proposed Policy. (A)Closed-Set 3D Object Detection denotes the canonical pipeline for clustering-based [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 4.** Building upon these observations, we choose PBNet [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Figures from the paper (1 more)

**Figure 5.** Figure 5: Experiments with System II. several additional key findings from our experiments as follows: Key Finding 1: The geometric structural relationship is vital. Minimizing reliance on visual deep learning features and emphasizing the geometric properties of point clouds ca…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

Towards Training-Free Open-World Classification with 3D Generative Models
cs.CV 2025-01 conditional novelty 6.0 of 10

A training-free 3D classifier creates anchor point clouds with text-to-3D generation and labels test objects by cosine similarity in a rotation-invariant feature space.

Reference graph

Works this paper leans on

51 extracted references · 46 canonical work pages · cited by 1 Pith paper

[1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...
[2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...
[3]

Ainetter, S.; and Fraundorfer, F. 2021. End-to-end Trainable Deep Neural Network for Robotic Grasp Detection and Semantic Segmentation from RGB . In International Conference on Robotics and Automation (ICRA), 13452--13458

work page 2021
[4]

Bicchi, A.; and Kumar, V. 2000. Robotic Grasping and Contact: A Review. In International Conference on Robotics and Automation (ICRA), 348--353

work page 2000
[5]

Bjorck, J.; Casta \ n eda, F.; Cherniadev, N.; Da, X.; Ding, R.; Fan, L.; Fang, Y.; Fox, D.; Hu, F.; Huang, S.; et al. 2025. Gr00t n1: An open foundation model for generalist humanoid robots. arXiv preprint arXiv:2503.14734

arXiv 2025
[6]

Bohg, J.; Morales, A.; Asfour, T.; and Kragic, D. 2014. Data-Driven Grasp Synthesis - A Survey. IEEE Transactions on Robotics (T-RO), 30(2): 289--309

work page 2014
[7]

Chen, S.; Fang, J.; Zhang, Q.; Liu, W.; and Wang, X. 2021. Hierarchical aggregation for 3d instance segmentation. In International Conference on Computer Vision (ICCV), 15467--15476

work page 2021
[8]

Cheng, A.-C.; Ji, Y.; Yang, Z.; Gongye, Z.; Zou, X.; Kautz, J.; B y k, E.; Yin, H.; Liu, S.; and Wang, X. 2025. Navila: Legged robot vision-language-action model for navigation. In Proceedings of Robotics: Science and Systems (RSS)

work page 2025

Show all 51 references

[9]

Cheng, X.; Li, J.; Yang, S.; Yang, G.; and Wang, X. 2024. Open-television: Teleoperation with immersive active visual feedback. In Conference on Robot Learning (CoRL)

2024
[10]

X.; Savva, M.; Halber, M.; Funkhouser, T.; and Nie ner, M

Dai, A.; Chang, A. X.; Savva, M.; Halber, M.; Funkhouser, T.; and Nie ner, M. 2017. ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. In Conference on Computer Vision and Pattern Recognition (CVPR)

2017
[11]

Dong, S.; Lin, G.; and Hung, T.-Y. 2022. Learning Regional Purity for Instance Segmentation on 3D Point Clouds. In European Conference on Computer Vision (ECCV), 56--72

2022
[12]

R.; Tian, S.; Yuan, W.; Krishna, R.; Fox, D.; Mandlekar, A.; and Guo, Y

Duan, J.; Pumacay, W.; Kumar, N.; Wang, Y. R.; Tian, S.; Yuan, W.; Krishna, R.; Fox, D.; Mandlekar, A.; and Guo, Y. 2024. AHA: A vision-language-model for detecting and reasoning over failures in robotic manipulation. arXiv preprint arXiv:2410.00371

2024 arXiv
[13]

A.; and Tripicchio, P

D’Avella, S.; Avizzano, C. A.; and Tripicchio, P. 2023. ROS-Industrial based robotic cell for Industry 4.0: Eye-in-hand stereo camera and visual servoing for flexible, fast, and accurate picking and hooking in the production line. Robotics and Computer-Integrated Manufacturing...

2023
[14]

Ekrem, \"O .; and Aksoy, B. 2023. Trajectory planning for a 6-axis robotic arm with particle swarm optimization algorithm. Engineering Applications of Artificial Intelligence (EAAI), 122: 106099

2023
[15]

Fang, H.-S.; Wang, C.; Gou, M.; and Lu, C. 2020. Graspnet-1billion: A large-scale benchmark for general object grasping. In Conference on Computer Vision and Pattern Recognition (CVPR), 11444--11453

2020
[16]

Fu, K.; and Dang, X. 2024. Light-weight convolutional neural networks for generative robotic grasping. IEEE Transactions on Industrial Informatics (T-II), 20(4): 6696--6707

2024
[17]

Graham, B. 2015. Sparse 3D Convolutional Neural Networks. In British Machine Vision Conference (BMVC), 150--159

2015
[18]

He, T.; Shen, C.; and Van Den Hengel, A. 2021. Dyco3d: Robust instance segmentation of 3d point clouds through dynamic convolution. In Conference on Computer Vision and Pattern Recognition (CVPR), 354--363

2021
[19]

Hirose, N.; Shah, D.; Sridhar, A.; and Levine, S. 2023. ExAug: Robot-conditioned navigation policies via geometric experience augmentation. In International Conference on Robotics and Automation (ICRA), 4077--4084

2023
[20]

Hoang, D.-C.; Nguyen, A.-N.; Vu, V.-D.; Nguyen, T.-U.; Vu, D.-Q.; Ngo, P.-Q.; Hoang, N.-A.; Phan, K.-T.; Tran, D.-T.; Nguyen, V.-T.; et al. 2024. Graspability-aware object pose estimation in cluttered scenes. IEEE Robotics and Automation Letters (RA-L), 9(4): 3124--3130

2024
[21]

Hu, W.; Lin, W.; Fang, H.; Wang, Y.; and Luo, D. 2024. OW3Det: Toward Open-World 3D Object Detection for Autonomous Driving. In International Conference on Intelligent Robots and Systems (IROS), 11587--11593

2024
[22]

Jiang, L.; Zhao, H.; Shi, S.; Liu, S.; Fu, C.-W.; and Jia, J. 2020. Pointgroup: Dual-set Point Grouping for 3D Instance Segmentation. In Conference on Computer Vision and Pattern Recognition (CVPR), 4867--4876

2020
[23]

Kim, S.; Lee, K.-H.; Kim, C.; and Yoon, J. 2024. Vision-centric 3D point cloud technique and custom gripper process for parcel depalletisation. The Journal of Intelligent Manufacturing (JIM), 1--17

2024
[24]

Li, B.; Ouyang, W.; Sheng, L.; Zeng, X.; and Wang, X. 2019. GS3D: An Efficient 3D Object Detection Framework for Autonomous Driving. In Conference on Computer Vision and Pattern Recognition (CVPR), 1019--1028

2019
[25]

Liang, Z.; Li, Z.; Xu, S.; Tan, M.; and Jia, K. 2021. Instance segmentation in 3d scenes using semantic superpoint tree networks. In International Conference on Computer Vision (ICCV), 2783--2792

2021
[26]

Liu, J.; Li, C.; Wang, G.; Lee, L.; Zhou, K.; Chen, S.; Xiong, C.; Ge, J.; Zhang, R.; and Zhang, S. 2024. Self-corrected multimodal large language model for end-to-end robot manipulation. arXiv preprint arXiv:2405.17418

2024 arXiv
[27]

Liu, J.; Zhang, R.; Fang, H.-S.; Gou, M.; Fang, H.; Wang, C.; Xu, S.; Yan, H.; and Lu, C. 2023. Target-referenced reactive grasping for dynamic objects. In Conference on Computer Vision and Pattern Recognition (CVPR), 8824--8833

2023
[28]

Luo, Z.; Cao, J.; Christen, S.; Winkler, A.; Kitani, K.; and Xu, W. 2024. Omnigrasp: Grasping diverse objects with simulated humanoids. In Conference on Neural Information Processing Systems (NeurIPS), 2161--2184

2024
[29]

R.; Litany, O.; He, K.; and Guibas, L

Qi, C. R.; Litany, O.; He, K.; and Guibas, L. J. 2019. Deep hough voting for 3d object detection in point clouds. In International Conference on Computer Vision (ICCV), 9277--9286

2019
[30]

R.; Su, H.; Mo, K.; and Guibas, L

Qi, C. R.; Su, H.; Mo, K.; and Guibas, L. J. 2017. Pointnet: Deep learning on point sets for 3d classification and segmentation. In Conference on Computer Vision and Pattern Recognition (CVPR), 652--660

2017
[31]

Tang, C.; Huang, D.; Dong, W.; Xu, R.; and Zhang, H. 2025. Foundationgrasp: Generalizable task-oriented grasping with foundation models. IEEE Transactions on Automation Science and Engineering (T-ASE)

2025
[32]

Thalhammer, S.; Bauer, D.; H \"o nig, P.; Weibel, J.-B.; Garc \' a-Rodr \' guez, J.; and Vincze, M. 2024. Challenges for monocular 6-d object pose estimation in robotics. IEEE Transactions on Robotics (T-RO), 40: 4065--4084

2024
[33]

Tong, L.; Song, K.; Tian, H.; Man, Y.; Yan, Y.; and Meng, Q. 2024. A novel RGB-D cross-background robot grasp detection dataset and background-adaptive grasping network. IEEE Transactions on Instrumentation and Measurement (T-IM)

2024
[34]

M.; Nguyen, T.; and Yoo, C

Vu, T.; Kim, K.; Luu, T. M.; Nguyen, T.; and Yoo, C. D. 2022. Softgroup for 3d instance segmentation on point clouds. In Conference on Computer Vision and Pattern Recognition (CVPR), 2708--2717

2022
[35]

Wang, R.; Zhang, J.; Chen, J.; Xu, Y.; Li, P.; Liu, T.; and Wang, H. 2023. DexGraspNet: A Large-Scale Robotic Dexterous Grasp Dataset for General Objects Based on Simulation. In International Conference on Robotics and Automation (ICRA), 11359--11366

2023
[36]

Wu, X.; Jiang, L.; Wang, P.-S.; Liu, Z.; Liu, X.; Qiao, Y.; Ouyang, W.; He, T.; and Zhao, H. 2024. Point transformer v3: Simpler faster stronger. In Conference on Computer Vision and Pattern Recognition (CVPR), 4840--4851

2024
[37]

Wu, Y.; Shi, M.; Du, S.; Lu, H.; Cao, Z.; and Zhong, W. 2022. 3D Instances as 1D Kernels. In European Conference on Computer Vision (ECCV), 235--252

2022
[38]

Xia, X.; Zhao, W.; Yan, Y.; Yang, G.; Zhang, R.; Huang, K.; and Yang, X. 2025. Towards Training-Free Open-World Classification with 3D Generative Models. In ACM International Conference on Multimedia (ACM MM)

2025
[39]

K.; Li, Z.; and Zhao, H

Xu, Z.; Zhang, Y.; Xie, E.; Zhao, Z.; Guo, Y.; Wong, K.-Y. K.; Li, Z.; and Zhao, H. 2024. Drivegpt4: Interpretable end-to-end autonomous driving via large language model. IEEE Robotics and Automation Letters (RA-L)

2024
[40]

Yin, Y.; Liu, Y.; Xiao, Y.; Cohen-Or, D.; Huang, J.; and Chen, B. 2024. Sai3d: Segment any instance in 3d scenes. In Conference on Computer Vision and Pattern Recognition (CVPR), 3292--3302

2024
[41]

Yin, Z.-H.; Wang, C.; Pineda, L.; Hogan, F.; Bodduluri, K.; Sharma, A.; Lancaster, P.; Prasad, I.; Kalakrishnan, M.; Malik, J.; et al. 2025. DexterityGen: Foundation Controller for Unprecedented Dexterity. In Proceedings of Robotics: Science and Systems (RSS)

2025
[42]

Ze, Y.; Zhang, G.; Zhang, K.; Hu, C.; Wang, M.; and Xu, H. 2024. 3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations. In Proceedings of Robotics: Science and Systems (RSS)

2024
[43]

Zhang, G.; Chen, J.; Gao, G.; Li, J.; Liu, S.; and Hu, X. 2024. Safdnet: A simple and effective network for fully sparse 3d object detection. In Conference on Computer Vision and Pattern Recognition (CVPR), 14477--14486

2024
[44]

Zhao, W.; Yan, Y.; Yang, C.; Ye, J.; Yang, X.; and Huang, K. 2023. Divide and conquer: 3d point cloud instance segmentation with point-wise binarization. In International Conference on Computer Vision (ICCV), 562--571

2023
[45]

Zhao, W.; Yang, G.; Zhang, R.; Jiang, C.; Yang, C.; Yan, Y.; Hussain, A.; and Huang, K. 2025 a . Open-pose 3d zero-shot learning: Benchmark and challenges. Neural Networks (NN), 181: 106775

2025
[46]

Zhao, W.; Zhang, R.; Wang, Q.; Cheng, G.; and Huang, K. 2025 b . BFANet: Revisiting 3D Semantic Segmentation with Boundary Feature Analysis. In Conference on Computer Vision and Pattern Recognition (CVPR), 29395--29405

2025
[47]

Zheng, Y.; Chen, X.; Zheng, Y.; Gu, S.; Yang, R.; Jin, B.; Li, P.; Zhong, C.; Wang, Z.; Liu, L.; et al. 2024. Gaussiangrasper: 3d language gaussian splatting for open-vocabulary robotic grasping. IEEE Robotics and Automation Letters (RA-L)

2024
[48]

Zhong, M.; Chen, X.; Chen, X.; Zeng, G.; and Wang, Y. 2022. Maskgroup: Hierarchical point grouping and masking for 3d instance segmentation. In ICME, 1--6

2022
[49]

Zhong, Y.; Jiang, Q.; Yu, J.; and Ma, Y. 2025. Dexgrasp anything: Towards universal robotic dexterous grasping with physics awareness. In Conference on Computer Vision and Pattern Recognition (CVPR), 22584--22594

2025
[50]

Zhuang, C.; Li, S.; and Ding, H. 2023. Instance segmentation based 6D pose estimation of industrial objects using point clouds for robotic bin-picking. Robotics and Computer-Integrated Manufacturing (RCIM), 82: 102541

2023
[51]

Zhuang, C.; Wang, H.; and Ding, H. 2024. AttentionVote: A coarse-to-fine voting network of anchor-free 6D pose estimation on point cloud for robotic bin-picking application. Robotics and Computer-Integrated Manufacturing (RCIM), 86: 102671

2024

Pith tools

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

[1] [1]

, " * write output.state after.block = add.period write newline

ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

[2] [2]

write newline

" write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

[3] [3]

Ainetter, S.; and Fraundorfer, F. 2021. End-to-end Trainable Deep Neural Network for Robotic Grasp Detection and Semantic Segmentation from RGB . In International Conference on Robotics and Automation (ICRA), 13452--13458

work page 2021

[4] [4]

Bicchi, A.; and Kumar, V. 2000. Robotic Grasping and Contact: A Review. In International Conference on Robotics and Automation (ICRA), 348--353

work page 2000

[5] [5]

Bjorck, J.; Casta \ n eda, F.; Cherniadev, N.; Da, X.; Ding, R.; Fan, L.; Fang, Y.; Fox, D.; Hu, F.; Huang, S.; et al. 2025. Gr00t n1: An open foundation model for generalist humanoid robots. arXiv preprint arXiv:2503.14734

arXiv 2025

[6] [6]

Bohg, J.; Morales, A.; Asfour, T.; and Kragic, D. 2014. Data-Driven Grasp Synthesis - A Survey. IEEE Transactions on Robotics (T-RO), 30(2): 289--309

work page 2014

[7] [7]

Chen, S.; Fang, J.; Zhang, Q.; Liu, W.; and Wang, X. 2021. Hierarchical aggregation for 3d instance segmentation. In International Conference on Computer Vision (ICCV), 15467--15476

work page 2021

[8] [8]

Cheng, A.-C.; Ji, Y.; Yang, Z.; Gongye, Z.; Zou, X.; Kautz, J.; B y k, E.; Yin, H.; Liu, S.; and Wang, X. 2025. Navila: Legged robot vision-language-action model for navigation. In Proceedings of Robotics: Science and Systems (RSS)

work page 2025