Pith. sign in

REVIEW 6 major objections 7 minor 59 references

Enhancing Human-Robot Collaboration: A Sim2Real Domain Adaptation Algorithm for Point Cloud Segmentation in Industrial Environments

T0 review · 6 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read FUSION, a dual DGCNN and residual-CNN network, claims to close the sim-to-real gap for industrial point cloud segmentation, reaching about 97.8 percent real-world accuracy after fine-tuning on only 25 real frames.

desk verdict A useful dataset and honest limitations, but the central SOTA accuracy claim is undercut by an underspecified evaluation protocol and conflicting numbers. read the letter →

arxiv 2506.09552 v1 pith:2QUHTLSH submitted 2025-06-11 cs.RO cs.CV

Fatemeh Mohammadi Amin , Darwin G. Caldwell , Hans Wernher van de Venn This is my paper
classification cs.ROcs.CV
keywords sim-to-realdomainadaptationpointcloudsemanticsegmentationhuman-robotcollaborationFUSIONnetworkDGCNNLiDARindustrialenvironmentCOVEREDdataset
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 tries to establish that a dual-stream network called FUSION, combining a dynamic graph convolutional network (DGCNN) with a residual convolutional network, can transfer semantic segmentation of industrial LiDAR point clouds from simulation to real human-robot collaboration scenes using only 25 annotated real frames for fine-tuning. If true, it matters because real industrial annotation is the bottleneck: a model could be trained mostly on flexible synthetic data and calibrated with a tiny real sample, while keeping the accuracy needed to identify humans, robots, and vehicles for safety. The authors report 98.9 percent accuracy on simulation data, about 45 percent when the simulation-only model is dropped into the real world, and about 97.9 percent after fine-tuning, with real-time inference near 0.1 seconds per frame after downsampling and static-object exclusion.

What carries the argument

The central object is the dual-stream FUSION architecture plus the frozen-feature fine-tuning scheme. In the DGCNN stream, the EdgeConv operation computes for each point a feature from its k nearest neighbors via $h'(i) = \max_{j:(i,j)\in E} \Theta \cdot (h_i \oplus (h_j - h_i))$, rebuilding the graph each layer; in the CNN-residual stream, each residual block computes $y = \mathrm{ReLU}(\mathrm{Conv2d}(\mathrm{ReLU}(\mathrm{Conv2d}(x)))) + x$, so gradients can bypass the convolutions. The pairing lets one stream preserve local geometric structure while the other learns multi-level appearance features, and freezing all but the last two fully connected layers during real-data fine-tuning preserves the simulation-learned representation while adapting the classifier.

What would settle it

Use the released simulation and real datasets, fix an 81-frame test split from distinct lab sessions and 25 fine-tuning frames from different sessions, and train RandLaNet, PointNet, DGCNN, CNN-Residual, and FUSION with identical pretraining and identical last-layer fine-tuning; if any baseline reaches or exceeds FUSION's 97.76 percent overall accuracy and 0.954 mIoU under that protocol, the paper's central superiority claim would be refuted.

Watch

Extended reading notes

Core claim

On its own terms, the paper's discovery is that the sim-to-real drop in this industrial HRC setting is largely recoverable by a specific architecture and a specific fine-tuning recipe. The architecture is FUSION, whose DGCNN stream extracts local geometry through dynamically recomputed k-nearest-neighbor edge features while the CNN-residual stream supplies hierarchical features with residual connections. The recipe freezes all layers except the last two fully connected layers and retrains them on 25 real point clouds. The authors report that this raises accuracy from 45 percent to 97.92 percent on the seven-class real test set (97.67 percent with an eighth background class), with overall accuracy 97.76 percent, per-class accuracy 96.82 percent, and mean IoU 0.954, and that FUSION exceeds the compared RandLaNet, PointNet, DGCNN, and CNN-Residual baselines on most classes while running faster than RandLaNet.

Load-bearing premise

The state-of-the-art claim rests on the assumption that every comparison network received the same simulation pretraining and the same 25-sample fine-tuning protocol, and that the 81 test point clouds are independent of the fine-tuning set; if the split or protocol differed across methods, the accuracy gap could come from unequal training rather than the architecture.

Editorial extensions

If this is right

  • Deploying FUSION in a new industrial cell would require only about 25 labeled real frames rather than a large annotated LiDAR dataset.
  • The fine-tuning recipe (freeze backbone, retrain last layers) should transfer to other point cloud backbones, offering a cheap adaptation baseline for HRC segmentation.
  • At roughly 0.1 seconds per frame after voxel downsampling and static-object exclusion, the model is fast enough for near-real-time safety monitoring in collaborative workspaces.
  • The extended COVERED dataset and its paired synthetic set give the community a common testbed for Sim2Real industrial point cloud segmentation.

Reading between the lines

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

  • If the reported jump is mainly a classifier-calibration effect, then even lighter adaptation—for instance, re-estimating batch-normalization statistics or training a small logistic head on frozen features—might reproduce most of the 52-point gain; that is a cheap experiment the paper does not run.
  • Because the simulation could not render dynamic objects and used a humanoid stand-in for humans, the 97.9 percent figure likely applies to quasi-static scenes; testing on sequences with articulated human motion would reveal the real safety-relevant limit.
  • The per-class results suggest FUSION excels on humans and robots, the safety-critical classes, while RandLaNet is slightly better on rigid AGVs; a natural extension is blending RandLaNet's multiscale sampling with FUSION's feature streams.
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

6 major / 7 minor

Summary. The paper proposes FUSION, a dual-stream network combining a Dynamic Graph Convolutional Neural Network (DGCNN) with a CNN augmented by residual connections, for semantic segmentation of 3D point clouds in industrial human-robot collaboration (HRC) settings. The intended use is Sim2Real domain adaptation: the network is pretrained on about 300 synthetic IsaacSim point clouds and then fine-tuned on 25 real point clouds from the authors' COVERED dataset. The authors report roughly 97.76% overall accuracy and 0.954 mIoU on real-world point clouds, claim state-of-the-art performance, and describe real-time deployment at about 0.098 s per frame. The paper also introduces an updated version of the COVERED dataset with additional classes and samples.

Significance. If the quantitative claims are fully supported, the paper would make a useful applied contribution: a concrete Sim2Real recipe for point-cloud segmentation in HRC, a partially public real-world dataset, and a demonstration of a fine-tuned architecture reaching high accuracy on real industrial scenes. The fine-tuning strategy (pretrain on simulation, freeze all but the last layers, fine-tune on 25 real samples) is simple and potentially practical. However, the manuscript currently does not provide enough architectural detail to reproduce FUSION, does not document the experimental protocol for baselines and data splits, reports inconsistent accuracy numbers without error bars, and evaluates real-time operation only in terms of timing. The central claim of improved state-of-the-art performance is therefore not yet substantiated. The authors should be credited for publishing the dataset and for testing in physical laboratory setups, but the evaluation needs substantial strengthening before the accuracy and SOTA claims can be accepted.

major comments (6)
  1. [Section 3.3, Equations (1)-(4), Figure 3] The architecture description is not reproducible. The CNN-residual stream is said to use 2D convolutional layers, but the input is a 3D point cloud; no projection, voxelization, multi-view rendering, or other transformation is defined. The fusion mechanism of the "dual-stream" network is also not described: there is no statement of where the DGCNN and CNN-residual features are combined, how the fused representation is decoded into per-point semantic labels, or what the final segmentation head looks like. Equation (1) defines h(p,q)=phi(p-q), but Equation (2) uses h_i and h_j in the EdgeConv update without defining these feature vectors or their relationship to Equation (1). Additionally, Section 4.2 mentions freezing "the last two fully connected layers," which is unclear for a convolutional segmentation architecture. Please provide a complete layer-by-layer specification, including tensor shapes, the exact fusion operation, and the loss/head details.
  2. [Section 4.2 and Table 2] The reported accuracies are inconsistent and need reconciliation. The abstract and Table 2 report 97.76% overall accuracy; Section 4.2 states 97.92% accuracy for the seven-class evaluation and 97.67% for the complete set of eight classes; Table 2 also gives per-class accuracy 96.82% and mIOU 0.954. It is not clear which number corresponds to which metric and which class set, nor why the Table 2 value does not match either of the Section 4.2 values. No error bars, confidence intervals, or multiple-seed results are provided despite only 81 test point clouds; the differences among methods in Table 2 (e.g., 97.76 vs. 96.34) are small enough that statistical significance must be established.
  3. [Section 4.1] The data split protocol is underspecified, which directly affects the validity of the central accuracy claim. The paper states that 80% of data was used for training and 20% for evaluation, and separately that 25 samples were used for fine-tuning and 81 for final testing. The authors must state whether the 25+81=106 samples constitute the 20% held-out portion, how the split was performed (random, chronological, scene-disjoint, or sensor-disjoint), whether the fine-tuning set is strictly disjoint from the test set, and how many real samples were available in total. Without this information, the 97.76% result could be inflated by temporal or scene leakage, or by fine-tuning on samples that are overly similar to the test set.
  4. [Section 4, Tables 1-2] The baseline comparison is not documented well enough to support the state-of-the-art claim. The paper says "A fair comparison is ensured by standardizing the testing conditions," but it does not specify whether RandLaNet, PointNet, DGCNN, and CNN-Residual were pretrained on the same 300 synthetic clouds, fine-tuned on exactly the same 25 real samples, and evaluated with the same frozen-layer policy, optimizer, learning-rate schedule, augmentation, point count (11,000), and KNN parameter. Unequal fine-tuning or different pretraining schedules could explain the reported margin. In addition, the comparison includes only generic segmentation networks; no domain-adaptation baselines (e.g., ePointDA, Complete & Label, or the methods discussed in Section 2.3) are evaluated, so the claim of improved "state-of-the-art" performance in Sim2Real domain adaptation is not established.
  5. [Section 4.4] The real-time evaluation reports only timing and qualitative visualization. The text gives prediction times of about 0.19 s, then 0.098 s after optimization, with segmentation and preprocessing taking 0.073 s, but it does not report segmentation accuracy, IoU, or any quantitative quality metric for the real-time runs. It is also not clear how the 0.073 s figure relates to the 0.115 s inference time reported for FUSION in Table 2. Please provide quantitative segmentation quality on the live sequences (or clarify that the offline test data are the same as the real-time data) and reconcile the timing discrepancy.
  6. [Section 6 and Section 3.1] The paper's own limitations section states that IsaacSim "could not render the point cloud for dynamic objects" and that "we utilized a humanoid object as a surrogate for a real human." This directly affects the source-domain representation of the Human class, which is central to the claimed HRC safety contribution. Please explain how the domain gap induced by using a static humanoid surrogate is mitigated, or temper the claim that the method achieves accurate human segmentation in real dynamic HRC scenarios.
minor comments (7)
  1. [Table 1] The RandLaNet row contains a formatting error: "0.8800.992" should presumably be "0.880, 0.992".
  2. [Section 4.2 and Table 2] The class-set description is confusing: Table 2 lists seven classes, while the text refers to a complete set of eight classes including "Unlabeled." Please state explicitly which classes are used in each reported metric and whether "Unlabeled" is evaluated or excluded.
  3. [Section 3.1 and Section 4.1] The paper should report basic dataset statistics for the updated COVERED dataset and the synthetic dataset: total number of real samples, per-class point distribution, number of LiDAR sequences or scenes, and how the 300 synthetic clouds are split across scenarios. This is needed to interpret the 80/20 split and the 25-sample fine-tuning set.
  4. [References] Several references are incomplete or malformed (e.g., [42], [51], [54], [55] lack full titles and venues), and the reference list should be cleaned up before publication.
  5. [Figure 3] Figure 3 is difficult to read and does not show tensor dimensions or the fusion point. Please provide a higher-resolution diagram with explicit feature-map sizes and the fusion operation labeled.
  6. [Data Availability] The Data Availability statement says the simulation dataset and code "will be made publicly available upon acceptance," but the abstract claims the dataset is published. Please clarify the current availability status of the code, the synthetic dataset, and the updated COVERED annotations.
  7. [General] The manuscript contains numerous language and typographical issues (e.g., "The Overal concept" in Figure 1, "this models" in Section 4.1, inconsistent spacing in tables). A careful proofreading pass is recommended.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the headline accuracy is an empirical measurement on a published dataset, not a quantity forced by construction or by self-citation.

full rationale

The paper's central claim is a measured accuracy (97.76% in Table 2; 97.92%/97.67% in Section 4.2) obtained by training FUSION on synthetic data, fine-tuning on 25 real samples, and testing on 81 real samples. No equation in the paper defines these accuracies as functions of the training labels or of the comparison baselines; the result is an experimental observation, not a derived quantity. The main self-citation is to the authors' COVERED dataset [50], which is publicly available and used as the target domain; citing one's own dataset is normal attribution and does not make the evaluation circular, because the measured IoU and accuracy values are not entailed by the dataset's construction. The comparison with RandLaNet, PointNet, DGCNN, and CNN-Residual under a standardized protocol (Section 4) gives independent content to the state-of-the-art claim. The paper's limitations (Section 6: IsaacSim could not render dynamic objects, humanoid surrogate used) and the underspecified 80/20 and 25/81 split are reproducibility and validity risks, not self-definitional reductions. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no known result is merely relabeled. Accordingly, no circular step can be quoted with a specific reduction.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the assumption that IsaacSim synthetic point clouds are adequate stand-ins for real industrial scenes, that fine-tuning with frozen layers is a valid adaptation strategy, and that the dual-stream fusion provides a benefit. The free parameters (k, point cloud size, downsampling rates, learning rate, fine-tuning sample count) are manually chosen without sensitivity analysis.

free parameters (5)
  • k in KNN = 10
    Chosen by testing on COVERED; lower k (10) outperformed k=20.
  • point cloud size = 11000
    Downsampled to 11000 points due to system capacity.
  • downsampling rates = 10:1 static, 2:1 dynamic
    Static objects downsampled more aggressively to preserve dynamic detail.
  • learning rate = 0.001 with cosine decay
    Set initially to 0.001.
  • fine-tuning sample count = 25
    25 real samples used for fine-tuning, a number chosen without justification.
assumptions (3)
  • domain assumption IsaacSim synthetic point clouds represent real industrial HRC domains well enough for fine-tuning to transfer.
    Section 3.1 uses IsaacSim with a humanoid surrogate because it cannot render dynamic objects; the limitation is acknowledged.
  • domain assumption Freezing all but the last two fully connected layers during fine-tuning preserves learned knowledge while adapting.
    Section 4.2 states this design without ablations or validation.
  • ad hoc to paper Combining DGCNN and CNN-residual in parallel improves segmentation accuracy.
    Section 3.3 gives no ablation or theoretical justification; the fusion mechanism is not described.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Human-Robot Collaboration: A Sim2Real Domain Adaptation Algorithm for Point Cloud Segmentation in Industrial Environments." pith.science (2026). https://pith.science/paper/2QUHTLSH

@misc{pith2026250609552,
  author       = {Pith},
  title        = {Pith review of: Enhancing Human-Robot Collaboration: A Sim2Real Domain Adaptation Algorithm for Point Cloud Segmentation in Industrial Environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2QUHTLSH}},
  note         = {Machine review of arXiv:2506.09552}
}
read the original abstract

The robust interpretation of 3D environments is crucial for human-robot collaboration (HRC) applications, where safety and operational efficiency are paramount. Semantic segmentation plays a key role in this context by enabling a precise and detailed understanding of the environment. Considering the intense data hunger for real-world industrial annotated data essential for effective semantic segmentation, this paper introduces a pioneering approach in the Sim2Real domain adaptation for semantic segmentation of 3D point cloud data, specifically tailored for HRC. Our focus is on developing a network that robustly transitions from simulated environments to real-world applications, thereby enhancing its practical utility and impact on a safe HRC. In this work, we propose a dual-stream network architecture (FUSION) combining Dynamic Graph Convolutional Neural Networks (DGCNN) and Convolutional Neural Networks (CNN) augmented with residual layers as a Sim2Real domain adaptation algorithm for an industrial environment. The proposed model was evaluated on real-world HRC setups and simulation industrial point clouds, it showed increased state-of-the-art performance, achieving a segmentation accuracy of 97.76%, and superior robustness compared to existing methods.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

59 extracted references · 35 canonical work pages

  1. [1]

    International Journal of Modern Manufacturing Technologies, Special Issue, XI3, 103–109 (2019)

    Olender, M., Banas, W.: Cobots–future in production. International Journal of Modern Manufacturing Technologies, Special Issue, XI3, 103–109 (2019)

    work page 2019

  2. [2]

    Journal of Mechanical Design 143(4), 040802 (2021)

    Vicentini, F.: Collaborative robotics: a survey. Journal of Mechanical Design 143(4), 040802 (2021)

    work page 2021

  3. [3]

    Sensors20(21), 6347 (2020)

    Mohammadi Amin, F., Rezayati, M., Venn, H.W., Karimpour, H.: A mixed- perception approach for safe human–robot collaboration in industrial automation. Sensors20(21), 6347 (2020)

    work page 2020

  4. [4]

    Publications Office of the European Union (2020)

    Hamon, R., Junklewitz, H., Sanchez, I.: Robustness and explainability of artificial intelligence. Publications Office of the European Union (2020)

    work page 2020

  5. [5]

    In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp

    Milioto, A., Vizzo, I., Behley, J., Stachniss, C.: Rangenet ++: Fast and accurate lidar semantic segmentation. In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 4213–4220 (2019). https://doi.org/ 10.1109/IROS40897.2019.8967762

    arXiv 2019

  6. [6]

    International journal of multimedia information retrieval7(2), 87–93 (2018)

    Guo, Y., Liu, Y., Georgiou, T., Lew, M.S.: A review of semantic segmenta- tion using deep neural networks. International journal of multimedia information retrieval7(2), 87–93 (2018)

    work page 2018

  7. [7]

    Gao, B., Pan, Y., Li, C., Geng, S., Zhao, H.: Are we hungry for 3d lidar data for 17 semantic segmentation? CoRRabs/2006.04307(2020)

    work page Pith review arXiv 2020

  8. [8]

    A Survey on Deep Domain Adaptation for LiDAR Perception

    Triess, L.T., Dreissig, M., Rist, C.B., Z¨ ollner, J.M.: A survey on deep domain adaptation for lidar perception. CoRRabs/2106.02377(2021) 2106.02377

    work page Pith review arXiv 2021

Show all 59 references
  1. [9]

    CoRRabs/2004.03327(2020)

    Wang, X., Jr., M.H.A., Lee, G.H.: Cascaded refinement network for point cloud completion. CoRRabs/2004.03327(2020)

    2020 arXiv

  2. [10]

    Guo, Y., Wang, H., Hu, Q., Liu, H., Liu, L., Bennamoun, M.: Deep Learning for 3D Point Clouds: A Survey (2020)

    2020

  3. [11]

    In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp

    Yang, Z., Wang, L.: Learning relationships for multi-view 3d object recognition. In: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pp. 7504–7513 (2019). https://doi.org/10.1109/ICCV.2019.00760

    2019

  4. [12]

    CoRRabs/1703.08866(2017)

    Ma, L., St¨ uckler, J., Kerl, C., Cremers, D.: Multi-view deep learning for consistent semantic mapping with RGB-D cameras. CoRRabs/1703.08866(2017)

    2017 arXiv

  5. [13]

    In: 2019 International Conference on Robotics and Automation (ICRA), pp

    Wu, B., Zhou, X., Zhao, S., Yue, X., Keutzer, K.: Squeezesegv2: Improved model structure and unsupervised domain adaptation for road-object segmenta- tion from a lidar point cloud. In: 2019 International Conference on Robotics and Automation (ICRA), pp. 4376–4382 (2019). IEEE

    2019

  6. [14]

    In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp

    Milioto, A., Vizzo, I., Behley, J., Stachniss, C.: Rangenet++: Fast and accurate lidar semantic segmentation. In: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 4213–4220 (2019). IEEE

    2019

  7. [15]

    CoRRabs/1711.06396(2017)

    Zhou, Y., Tuzel, O.: Voxelnet: End-to-end learning for point cloud based 3d object detection. CoRRabs/1711.06396(2017)

    2017 arXiv

  8. [16]

    CoRRabs/1811.04337(2018)

    Meng, H., Gao, L., Lai, Y., Manocha, D.: Vv-net: Voxel V AE net with group convolutions for point cloud segmentation. CoRRabs/1811.04337(2018)

    2018 arXiv

  9. [17]

    CoRRabs/1908.06295(2019)

    Zhang, Z., Hua, B., Yeung, S.: Shellnet: Efficient point cloud convolutional neural networks using concentric shells statistics. CoRRabs/1908.06295(2019)

    2019 arXiv

  10. [18]

    CoRRabs/1802.04402(2018)

    Huang, Q., Wang, W., Neumann, U.: Recurrent slice networks for 3d segmentation on point clouds. CoRRabs/1802.04402(2018)

    2018 arXiv

  11. [19]

    In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    Landrieu, L., Simonovsky, M.: Large-scale point cloud semantic segmentation with superpoint graphs. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2018)

    2018

  12. [20]

    CoRRabs/1612.00593(2016)

    Qi, C.R., Su, H., Mo, K., Guibas, L.J.: Pointnet: Deep learning on point sets for 3d classification and segmentation. CoRRabs/1612.00593(2016)

    2016 arXiv

  13. [21]

    CoRRabs/1706.02413(2017) 18 1706.02413

    Qi, C.R., Yi, L., Su, H., Guibas, L.J.: Pointnet++: Deep hierarchical fea- ture learning on point sets in a metric space. CoRRabs/1706.02413(2017) 18 1706.02413

    2017 arXiv

  14. [22]

    CoRRabs/1801.07791(2018) 1801.07791

    Li, Y., Bu, R., Sun, M., Chen, B.: Pointcnn. CoRRabs/1801.07791(2018) 1801.07791

    2018 arXiv

  15. [23]

    CoRRabs/1801.07829(2018) 1801.07829

    Wang, Y., Sun, Y., Liu, Z., Sarma, S.E., Bronstein, M.M., Solomon, J.M.: Dynamic graph CNN for learning on point clouds. CoRRabs/1801.07829(2018) 1801.07829

    2018 arXiv

  16. [24]

    CoRRabs/1909.08291(2019)

    Aksoy, E.E., Baci, S., Cavdar, S.: Salsanet: Fast road and vehicle segmentation in lidar point clouds for autonomous driving. CoRRabs/1909.08291(2019)

    2019 arXiv

  17. [25]

    CoRRabs/1710.07368(2017)

    Wu, B., Wan, A., Yue, X., Keutzer, K.: Squeezeseg: Convolutional neural nets with recurrent CRF for real-time road-object segmentation from 3d lidar point cloud. CoRRabs/1710.07368(2017)

    2017 arXiv

  18. [26]

    IEEE Robotics and Automation Letters3(4), 3434–3440 (2018) https://doi.org/10.1109/LRA.2018

    Zeng, Y., Hu, Y., Liu, S., Ye, J., Han, Y., Li, X., Sun, N.: Rt3d: Real-time 3-d vehicle detection in lidar point cloud for autonomous driving. IEEE Robotics and Automation Letters3(4), 3434–3440 (2018) https://doi.org/10.1109/LRA.2018. 2852843

    2018 doi

  19. [27]

    CoRRabs/2003.03653(2020)

    Cortinhal, T., Tzelepis, G., Aksoy, E.E.: Salsanext: Fast semantic segmentation of lidar point clouds for autonomous driving. CoRRabs/2003.03653(2020)

    2020 arXiv

  20. [28]

    CoRRabs/2102.04530(2021)

    Cheng, R., Razani, R., Taghavi, E., Li, E., Liu, B.: (af)2-s3net: Attentive feature fusion with adaptive feature selection for sparse semantic segmentation network. CoRRabs/2102.04530(2021)

    2021 arXiv

  21. [29]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Thomas, H., Qi, C.R., Deschaud, J.-E., Marcotegui, B., Goulette, F., Guibas, L.J.: Kpconv: Flexible and deformable convolution for point clouds. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 6411–6420 (2019)

    2019

  22. [30]

    CoRR abs/1804.05132(2018)

    Feng, D., Rosenbaum, L., Dietmayer, K.: Towards safe autonomous driving: Cap- ture uncertainty in the deep neural network for lidar 3d vehicle detection. CoRR abs/1804.05132(2018)

    2018 arXiv

  23. [31]

    Applied Sciences10, 1183 (2020) https://doi.org/10.3390/app10031183

    Zha, Fu, J., Wang, Zhaoyu, G., Yin-Sheng, L., Yidong, C.: Semantic 3d recon- struction for robotic manipulators with an eye-in-hand vision system. Applied Sciences10, 1183 (2020) https://doi.org/10.3390/app10031183

    2020 doi

  24. [32]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

    Hu, Q., Yang, B., Xie, L., Rosa, S., Guo, Y., Wang, Z., Trigoni, N., Markham, A.: Randla-net: Efficient semantic segmentation of large-scale point clouds. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 11108–11117 (2020)

    2020

  25. [33]

    Robotics and Computer-Integrated Manu- facturing (2021) https://doi.org/10.1016/j.rcim.2021.102304

    Fan, J., Zheng, P., Li, S.: Vision-based holistic scene understanding towards 19 proactive human-robot collaboration. Robotics and Computer-Integrated Manu- facturing (2021) https://doi.org/10.1016/j.rcim.2021.102304

    2021

  26. [34]

    Robotics and Computer-Integrated Manufacturing86, 102663 (2024) https://doi.org/10.1016/ j.rcim.2023.102663

    Angleraud, A., Ekrekli, A., Samarawickrama, K., Sharma, G., Pieters, R.: Sensor-based human–robot collaboration for industrial tasks. Robotics and Computer-Integrated Manufacturing86, 102663 (2024) https://doi.org/10.1016/ j.rcim.2023.102663

    2024

  27. [35]

    Frontiers in Neurorobotics17 (2023) https://doi.org/10.3389/fnbot.2023.1084000

    Su, H., Qi, W., Chen, J., Yang, C., Sandoval, J., Laribi, M.A.: Recent advance- ments in multimodal human–robot interaction. Frontiers in Neurorobotics17 (2023) https://doi.org/10.3389/fnbot.2023.1084000

    2023

  28. [36]

    Advanced Engineering Informatics65, 103109 (2025) https://doi.org/10.1016/j.aei.2025

    Liu, J., Luo, H., Wu, D.: Human–robot collaboration in construction: Robot design, perception and interaction, and task allocation and execution. Advanced Engineering Informatics65, 103109 (2025) https://doi.org/10.1016/j.aei.2025. 103109

    2025 doi

  29. [37]

    The International Journal of Advanced Manufacturing Technology132, 1–23 (2024) https://doi.org/10.1007/s00170-024-13385-2

    Duan, J., Zhuang, L., Zhang, Q., Zhou, Y., Qin, J.: Multimodal perception- fusion-control and human–robot collaboration in manufacturing: a review. The International Journal of Advanced Manufacturing Technology132, 1–23 (2024) https://doi.org/10.1007/s00170-024-13385-2

    2024 doi

  30. [38]

    IEEE Robotics and Automation Letters9(5), 4186–4193 (2024) https: //doi.org/10.1109/LRA.2024.3376974

    Lin, J., Zhong, K., Gong, T., Zhang, X., Wang, N.: Prior information-assisted neural network for point cloud segmentation in human-robot interaction sce- narios. IEEE Robotics and Automation Letters9(5), 4186–4193 (2024) https: //doi.org/10.1109/LRA.2024.3376974

    2024

  31. [39]

    798–805 (2023)

    Alharasees, O., Adali, O.H., Kale, U.: Human factors in the age of autonomous uavs: Impact of artificial intelligence on operator performance and safety, pp. 798–805 (2023). https://doi.org/10.1109/ICUAS57906.2023.10156037

    2023

  32. [40]

    International Journal of Industrial Ergonomics104, 103663 (2024) https://doi.org/10.1016/j.ergon.2024

    de Nobile, A., Bibbo, D., Russo, M., Conforto, S.: A focus on quantitative meth- ods to assess human factors in collaborative robotics. International Journal of Industrial Ergonomics104, 103663 (2024) https://doi.org/10.1016/j.ergon.2024. 103663

    2024 doi

  33. [41]

    IEEE Transactions on Automation Science and EngineeringPP, 1–1 (2025) https://doi.org/10.1109/TASE.2025.3528051

    Chen, H., Li, S., Fan, J., Duan, A., Yang, C., Navarro-Alarcon, D., Zheng, P.: Human-in-the-loop robot learning for smart manufacturing: A human-centric per- spective. IEEE Transactions on Automation Science and EngineeringPP, 1–1 (2025) https://doi.org/10.1109/TASE.2025.3528051

    2025

  34. [42]

    https://doi.org/10.1561/2300000052

    Lasota, P.A., Song, T., Shah, J.A., (2017). https://doi.org/10.1561/2300000052

    2017 doi

  35. [43]

    Frontiers in Robotics and AI9(2022) https://doi.org/10.3389/frobt.2022.799522 20

    Hopko, S., Wang, J., Mehta, R.: Human factors considerations and metrics in shared space human-robot collaboration: A systematic review. Frontiers in Robotics and AI9(2022) https://doi.org/10.3389/frobt.2022.799522 20

    2022

  36. [44]

    Journal of Occupational Accidents9, 27–45 (1987)

    Jiang, B., Gainer, C.: A cause-and-effect analysis of robot accidents. Journal of Occupational Accidents9, 27–45 (1987)

    1987

  37. [45]

    In: 2020 IEEE International Conference on Robotics and Automation (ICRA), pp

    El-Shamouty, M., Wu, X., Yang, S., Albus, M., Huber, M.F.: Towards safe human-robot collaboration using deep reinforcement learning. In: 2020 IEEE International Conference on Robotics and Automation (ICRA), pp. 4899–4905 (2020). https://doi.org/10.1109/ICRA40945.2020.9196924

    2020

  38. [46]

    Sensors23(21) (2023) https://doi.org/ 10.3390/s23218989

    Amaral, P., Silva, F., Santos, V.: Recognition of grasping patterns using deep learning for human–robot collaboration. Sensors23(21) (2023) https://doi.org/ 10.3390/s23218989

    2023 doi

  39. [47]

    In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp

    Armeni, I., Sener, O., Zamir, A.R., Jiang, H., Brilakis, I., Fischer, M., Savarese, S.: 3d semantic parsing of large-scale indoor spaces. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 1534–1543 (2016). https://doi.org/10.1109/CVPR.2016.170

    2016 doi

  40. [48]

    CoRR abs/1704.03847(2017)

    Hackel, T., Savinov, N., Ladicky, L., Wegner, J.D., Schindler, K., Pollefeys, M.: Semantic3d.net: A new large-scale point cloud classification benchmark. CoRR abs/1704.03847(2017)

    2017 arXiv

  41. [49]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

    Behley, J., Garbade, M., Milioto, A., Quenzel, J., Behnke, S., Stachniss, C., Gall, J.: Semantickitti: A dataset for semantic scene understanding of lidar sequences. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 9297–9307 (2019)

    2019

  42. [50]

    In: 2022 IEEE 27th International Conference on Emerging Technologies and Factory Automation (ETF A), pp

    Munasinghe, C., Amin, F.M., Scaramuzza, D., Venn, H.W.: Covered, collabo- rative robot environment dataset for 3d semantic segmentation. In: 2022 IEEE 27th International Conference on Emerging Technologies and Factory Automation (ETF A), pp. 1–4 (2022). https://doi.org/10.1109...

    2022

  43. [51]

    CoRR (2020)

    Yi, L., Gong, B., Funkhouser, T.A.: Complete & label: A domain adaptation approach to semantic segmentation of lidar point clouds. CoRR (2020)

    2020

  44. [52]

    CoRRabs/2106.11239(2021)

    Jia, D., Hermans, A., Leibe, B.: Domain and modality gaps for lidar-based person detection on mobile robots. CoRRabs/2106.11239(2021)

    2021 arXiv

  45. [53]

    In: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2020)

    Langer, F., Milioto, A., Haag, A., Behley, J., Stachniss, C.: Domain transfer for semantic segmentation of lidar data using deep neural networks. In: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (2020). https://doi.org/10.1109/IROS45743.2020.9341508

    2020

  46. [54]

    CoRR (2020)

    Zhao, S., Wang, Y., Li, B., Wu, B., Gao, Y., Xu, P., Darrell, T., Keutzer, K.: epointda: An end-to-end simulation-to-real domain adaptation framework for lidar point cloud segmentation. CoRR (2020)

    2020

  47. [55]

    CoRR (2020)

    Jiang, P., Saripalli, S.: Lidarnet: A boundary-aware domain adaptation model for 21 lidar point cloud semantic segmentation. CoRR (2020)

    2020

  48. [56]

    Automation in Construction130, 103874 (2021) https://doi.org/10.1016/j.autcon.2021.103874

    Yin, C., Wang, B., Gan, V.J.L., Wang, M., Cheng, J.C.P.: Automated seman- tic segmentation of industrial point clouds using respointnet++. Automation in Construction130, 103874 (2021) https://doi.org/10.1016/j.autcon.2021.103874

    2021

  49. [57]

    Chen, Z., Xu, H., Chen, W., Zhou, Z., Xiao, H., Sun, B., Xie, X., Kang, W.: PointDC:Unsupervised Semantic Segmentation of 3D Point Clouds via Cross- modal Distillation and Super-Voxel Clustering (2024)

    2024

  50. [58]

    Wu, C., Bi, X., Pfrommer, J., Cebulla, A., Mangold, S., Beyerer, J.: Sim2real Transfer Learning for Point Cloud Segmentation: An Industrial Application Case on Autonomous Disassembly (2023)

    2023

  51. [59]

    Accessed: 01.02.2024

    NVIDIA: NVIDIA Isaac Sim. Accessed: 01.02.2024. https://developer.nvidia. com/isaac/sim Declarations 6.1 Funding This work was partially supported by DIZH (Digitalization Initiative of the Zurich Higher Education Institutions) funding. 6.2 Conflict of interest The authors have...

    2024

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.