Pith. sign in

REVIEW 3 major objections 6 minor 64 references

NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read Non-uniform radial voxels—spaced in an arithmetic progression—are claimed to make voxel-based LiDAR semantic segmentation both more accurate and several times cheaper in compute, memory, and training time than uniform cylindrical grids.

desk verdict Solid non-uniform partition mechanism with clean controlled ablations, undermined by an over-claimed SOTA statement and mixed-hardware efficiency comparisons; needs revision, not rejection. read the letter →

arxiv 2505.24634 v2 pith:PJYENZVW submitted 2025-05-30 cs.CV

classification cs.CV
keywords LiDARsemanticsegmentationnon-uniformpartitioncylindricalcoordinatesarithmeticprogressionofintervalsparseconvolutionKITTInuScenesvoxelrepresentation
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 argues that the standard way of dividing 3D space for LiDAR segmentation—uniform cylindrical voxels—is both wasteful and representational: equal radial intervals put too many voxels where points are dense and too few where they are sparse. The paper proposes NUC-Net, which spaces radial intervals in an arithmetic progression, so voxels grow slowly near the sensor and rapidly far away. At a quarter of the radial resolution of the standard Cylinder3D representation, the network reports higher semantic segmentation accuracy (73.6 vs 67.8 mIoU on SemanticKITTI test) with roughly three times faster inference, about four times less training time, and about half the GPU memory. If these comparisons hold under matched measurement, a simple change to how voxels are laid out, rather than a new network family, is a general efficiency lever for voxel-based LiDAR perception. The paper also gives a mechanism for the gain: smaller near-field voxels reduce label-mixing encoding errors, while larger far-field voxels widen the receptive field of fixed-size convolution kernels.

What carries the argument

The load-bearing object is the Arithmetic Progression of Interval (API) radial partition: the $i$-th radial interval is $a_i = a_0 + i\,d$, chosen so voxel volume grows approximately as $b h d^2 i^3/2$ rather than linearly in the radial index. It controls how many voxels are allocated to dense near-field vs sparse far-field regions, determining both encoding error and receptive field under a fixed 3D sparse-convolution backbone. A second component, non-uniform multi-scale aggregation, builds multi-scale voxel features whose radial intervals are aligned so that scale stacking does not create mismatched cell boundaries. The argument is carried by the volume-growth identity $V_{i,j,k}\propto i^3$ and the empirical point-density profile of LiDAR scans.

What would settle it

Run the identical 3D sparse-convolution network on SemanticKITTI with only the partition changed—uniform radial intervals vs API—at exactly the same voxel resolution, batch size, GPU, and inference script, and measure wall-clock latency, peak GPU memory, and validation mIoU. If the API partition does not reproduce a substantial mIoU gain at matched resolution (the paper's Table V reports +4.7 for Cylinder3D at $[120,360,32]$) and does not reduce non-empty voxels or active sites in deeper layers as claimed, the central efficiency-accuracy claim fails.

Watch

Extended reading notes

Core claim

The central claim is that replacing the uniform radial partition of cylindrical-coordinate voxelization with an Arithmetic Progression of Interval (API) partition—where the $i$-th radial interval is $a_i = a_0 + i d$—makes the voxel volume grow like $i^3$ instead of $i$, matching the way LiDAR point density falls off with distance. This yields a more balanced points-per-cell distribution, lower encoding error near the sensor, and a larger effective receptive field in distant regions without enlarging convolution kernels or raising resolution. The paper reports that this one change, applied as a general component, raises Cylinder3D from 61.8 to 66.5 mIoU and PolarNet from 52.6 to 58.7 mIoU at matched resolution on SemanticKITTI validation, and that NUC-Net reaches 73.6 mIoU on the SemanticKITTI test set at $120\times360\times32$ resolution with 19.9 GMACs and 56 ms inference. The same representation is shown to transfer to multi-scan semantic, panoptic, scribble-supervised, and corruption-robustness settings. The paper frames the contribution as a representation-level improvement that is complementary to efficient range-view, distillation, and lightweight-architecture methods.

Load-bearing premise

The reported speed, memory, and training-time gains assume the comparison numbers for the uniform baselines were measured under the same hardware and inference protocol as the paper's own runs; if Cylinder3D's 171 ms, PVKD's 76 ms, and other baseline timings come from different GPUs or settings, the claimed 3x, 4x, and 2x improvements shrink accordingly.

Editorial extensions

If this is right

  • If the central claim is correct, voxel-based LiDAR segmentation does not need higher input resolution to gain accuracy; the same accuracy can be reached at one-quarter the radial resolution.
  • Uniform cylindrical networks such as Cylinder3D and PolarNet can be upgraded to non-uniform partition without changing their architecture, gaining accuracy and speed simultaneously.
  • The reported 19.9 GMACs, 56 ms inference, and 2.5 GPU-days training profile would make state-of-the-art LiDAR segmentation practical on a single consumer GPU, lowering the barrier for real-time autonomous driving systems.
  • The non-uniform partition transfers to multi-scan, panoptic, and scribble-annotated settings, so the benefit is not specific to one task.
  • Because the method is representation-level, it can be combined with range-view, distillation, and lightweight-architecture approaches that target efficiency from other directions.

Reading between the lines

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

  • The volume-growth law $V\propto i^3$ suggests an explicit design rule the paper does not pursue: the radial interval schedule should track the inverse of the point-density profile of the sensor, so other sensors with different falloff rates could need different exponents or progressions, a testable per-sensor calibration.
  • The same API idea could be applied to the range (column) axis of spherical range-view representations, potentially transferring the speed-accuracy gain to 2D-convolution methods that currently dominate real-time deployments.
  • The paper's encoding-error argument predicts that non-uniform partition should help most in mixed-class boundary voxels near the sensor; a fine-grained evaluation of boundary IoU by distance bin would make that mechanism directly visible.
  • An implicit economic consequence is that if training memory indeed drops by roughly half, larger batch sizes or longer contexts such as multi-frame input become affordable without changing the model, extending the method's value beyond single-scan segmentation.
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

3 major / 6 minor

Summary. The paper proposes NUC-Net, a LiDAR semantic segmentation method that replaces the uniform radial partition of cylindrical voxel grids with a non-uniform Arithmetic Progression of Interval (API) partition, together with a non-uniform multi-scale aggregation module. The authors report state-of-the-art results on SemanticKITTI and nuScenes, with 4x faster training, 2x lower GPU memory, and 3x inference speedup relative to a uniform counterpart, and they provide a qualitative theoretical analysis based on encoding error and receptive-field arguments. The paper also includes generalization experiments to multi-scan segmentation, panoptic segmentation, scribble-annotated data, and corrupted-data robustness.

Significance. If the claims hold, the contribution is practically valuable: a partition change that keeps accuracy while allowing a 4x input-resolution reduction would make voxel-based LiDAR segmentation considerably cheaper. The strongest evidence is Table V, which shows clean, controlled gains from replacing uniform with non-uniform partition at identical voxel resolution for both PolarNet (+6.1 mIoU at [120,360,32]) and Cylinder3D (+4.7 mIoU at [120,360,32]). The breadth of additional experiments (multi-scan, panoptic, scribble, Robo3D) is a genuine strength, and the code is promised to be public. The theoretical analysis in Section IX is qualitative rather than formal, but the core empirical mechanism is supported by the controlled ablation.

major comments (3)
  1. [Section IV-C, Tables II and III] The state-of-the-art claim on nuScenes is internally contradicted by the paper's own numbers. On the nuScenes validation set, Table II reports SFPNet at 80.1 mIoU and NUC-Net at 78.9 mIoU; on the test set, Table III shows NUC-Net tied with SFPNet at 80.2 mIoU. SFPNet is a LiDAR-only, single-scan method and is therefore a direct comparison. The abstract and Section IV-C claim "state-of-the-art performance on SemanticKITTI and nuScenes datasets," which is not supported by these tables. The authors should revise the headline claim to "comparable" or "competitive" on nuScenes, or explicitly justify why SFPNet should be excluded from the comparison.
  2. [Abstract; Section IV-E, Tables I and XIV] The efficiency gains attributed to the non-uniform partition are not measured at matched resolution. Table I shows the uniform "Baseline" at the same input resolution (120x360x32) has 52 ms inference speed, while "Ours" has 56 ms, so the partition change alone does not produce a 3x inference speedup. The 3x/4x/2x figures in Table XIV compare NUC-Net at 120x360x32 against Cylinder3D at 480x360x32, conflating the effect of the partition with the effect of reducing input resolution. Since Table V reports only accuracy at matched resolution, the "general component" claim for efficiency is unsupported. Please report inference speed, GPU memory, and training time under matched resolution for the uniform and non-uniform versions, or qualify the claims to state that the efficiency gains come from the combination of non-uniform partition and reduced input resolution.
  3. [Section IV-D, Table IV vs Table I] The uniform baseline jumps from 63.2 mIoU on validation (Table IV) to 69.5 mIoU on test (Table I) without explanation. Because this baseline is the control for the headline SemanticKITTI comparison, the source of the discrepancy (e.g., test-time ensemble, instance augmentation, validation/test distribution) should be stated explicitly. Without this, the reader cannot reconcile the ablation numbers with the main results or assess how much of the final gain is attributable to the proposed components.
minor comments (6)
  1. [Throughout] There are several typos and inconsistencies: "wihch" (Section I), "V oxel" and "voxel-vise" in the Figure 2 caption, "Arthmetric" and "tollerance" in Section IV-D, "furhter" in Section VII, and "conprehensive" in Section VIII. Please proofread the manuscript.
  2. [Table I] The footnote marking asterisked speeds is insufficient for comparing efficiency claims. Cylinder3D's 171 ms and RPVNet's 168 ms are not marked as RTX 3090 measurements, and RangeFormer's 37 ms is explicitly from an A100. Please report the hardware and measurement protocol for every speed entry, not only the asterisked ones.
  3. [Table II] The row labeled "Salsanext [3]" cites reference [3] (PolarNet), which appears to be a citation error; the correct reference for SalsaNext should be used.
  4. [Table XIII and surrounding text] The voxel counts in Table XIII are non-integers (e.g., 21015.1), which suggests these are frame-averaged values; please state this explicitly. Also, "1.72 lower" should be "1.72x lower" (or "by a factor of 1.72").
  5. [Section VIII, Table XII] The text mentions Cylinder3D with input voxel resolution [120,360,180], but Table XII lists Cylinder3D[120,360,32]. Please align the text and table.
  6. [Section IX] The theoretical analysis is qualitative (encoding error, receptive field, sparse voxel communication). The abstract's phrase "theoretical analysis" overstates the level of formality; consider describing it as an empirical and qualitative analysis.

Circularity Check

1 steps flagged · score 4.0 of 10

Validation-set API gains are partly parameter-selection results, but the central claim retains independent test-set support.

  1. fitted input called prediction [Section IV-D, 'Different Settings for API'; Tables VI and VII]
    "After that, we select several parameter settings around a0 = 0.05 and train them individually on the training set. We then choose the parameter that achieves the best LiDAR point cloud segmentation performance on the validation set as the final parameter."

    The reported validation mIoU for NUC-Net (70.3 on SemanticKITTI, 78.9 on nuScenes) is the maximum over the candidate (a0, d) rows in Table VII, with the very same validation split used as the selection criterion. Therefore the validation-set improvements over the uniform baseline (4.4 mIoU in Table IV; 4.8 in Table VI) are resubstitution estimates: the 'prediction' is the objective being optimized. This does not invalidate the test-set results (73.6 SemanticKITTI test, 80.2 nuScenes test) or the matched-resolution gains in Table V, which provide independent evidence for the core claim.

full rationale

The paper's central mechanism, non-uniform cylindrical partition, is empirically evaluated rather than derived from a self-referential argument. The matched-resolution ablations in Table V show consistent gains when the API partition replaces uniform partition in PolarNet and Cylinder3D across multiple resolutions, and the SemanticKITTI/nuScenes test results are independent of the validation-based hyperparameter selection. The only circular element is the API parameter selection in Section IV-D: a0=0.05 and d=0.0062 are chosen as the best-performing candidates on the validation split, and the same split's mIoU is then reported as the method's validation performance. That makes the validation-set accuracy a selection result rather than an out-of-sample prediction. There is no load-bearing self-citation chain, no imported uniqueness theorem, and no definitional equivalence between the proposed representation and the reported SOTA claim. The abstract's unqualified 'state-of-the-art on nuScenes' is contradicted by the paper's own Table II (SFPNet 80.1 vs Ours 78.9), but that is a factual/correctness issue, not circularity.

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

The central gain is measured, not derived from a theory. The paper's contribution rests on four design choices (a_0, d, scale count, resolution), two of which (a_0, d) are explicitly tuned on the validation set and one (resolution) on ablation; additionally, the causal story relies on domain assumptions about LiDAR density profiles, dominant error sources, and sparse-convolution cost. The theoretical analysis is explanatory, not predictive, which is why the ledger lists assumptions rather than derived constants.

free parameters (4)
  • a_0 (first radial interval of API) = 0.05
    Selected in Section IV-D on the SemanticKITTI validation set; Table VI and Table VII show mIoU varies from 68.1 to 70.3 across nearby values, so the headline validation gain is partially tuned.
  • d (common difference of API intervals) = 0.0062
    Selected by the same validation-set procedure as a_0; together with a_0 the 120 intervals span about 50 m, which matches the effective evaluation range.
  • Aggregation scale count (t) = 2
    Reported in Section IV-B as 'The aggregation scale is set as 2'; no ablation is provided for this choice, so it is a hand-chosen design parameter.
  • Input voxel resolution (n_r, n_phi, n_z) = 120×360×32
    Chosen by ablation in Table VIII over [120,180,32], [120,360,32], and [480,360,32]; the central 4× resolution-reduction claim depends on this choice.
assumptions (4)
  • domain assumption LiDAR point density in SemanticKITTI peaks near 7 m and decays with range such that an i^3 voxel-volume growth better balances points per cell.
    Stated in Section IV-D and Fig. 4; used to motivate API parameters and to argue far-region encoding errors are negligible because most points are within 0-15 m (Section IX-A). If nuScenes' different sensor geometry breaks this profile, the transfer of the same (a_0, d) is unjustified.
  • domain assumption Within-voxel label mixing (encoding error) near the sensor and limited receptive field far from the sensor are the dominant error sources the partition should target.
    Section IX-A frames the entire analysis around encoding error and receptive field; the explanation presumes these two effects dominate over other error sources such as occlusion and class ambiguity, which is asserted rather than tested.
  • domain assumption Sparse-convolution cost is dominated by the number of non-empty voxels in shallow layers and by total voxel resolution in deep layers.
    Section IX-B follows Cylinder3D and Minkowski-style sparse convolution; this premise underlies the claimed 3× inference speedup and 4× training speedup, and it is not validated against profiler data in the paper.
  • standard math Standard asymptotic expansion used to obtain V_{i,j,k} approximately equal to (b h d^2 i^3)/2 from Eq. (3).
    The i^3 volume growth is derived under the approximation that a_0 is small and i is large; it is a heuristic scaling law, not an exact identity, and the paper uses it to argue the partition matches density decay.

how reviews work

0 comments
Cite this review

Pith. "Pith review of NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation." pith.science (2026). https://pith.science/paper/PJYENZVW

@misc{pith2026250524634,
  author       = {Pith},
  title        = {Pith review of: NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PJYENZVW}},
  note         = {Machine review of arXiv:2505.24634}
}
abstract

LiDAR semantic segmentation plays a vital role in autonomous driving. Existing voxel-based methods for LiDAR semantic segmentation apply uniform partition to the 3D LiDAR point cloud to form a structured representation based on cartesian/cylindrical coordinates. Although these methods show impressive performance, the drawback of existing voxel-based methods remains in two aspects: (1) it requires a large enough input voxel resolution, which brings a large amount of computation cost and memory consumption. (2) it does not well handle the unbalanced point distribution of LiDAR point cloud. In this paper, we propose a non-uniform cylindrical partition network named NUC-Net to tackle the above challenges. Specifically, we propose the Arithmetic Progression of Interval (API) method to non-uniformly partition the radial axis and generate the voxel representation which is representative and efficient. Moreover, we propose a non-uniform multi-scale aggregation method to improve contextual information. Our method achieves state-of-the-art performance on SemanticKITTI and nuScenes datasets with much faster speed and much less training time. And our method can be a general component for LiDAR semantic segmentation, which significantly improves both the accuracy and efficiency of the uniform counterpart by $4 \times$ training faster and $2 \times$ GPU memory reduction and $3 \times$ inference speedup. We further provide theoretical analysis towards understanding why NUC is effective and how point distribution affects performance. Code is available at \href{https://github.com/alanWXZ/NUC-Net}{https://github.com/alanWXZ/NUC-Net}.

Figures

Figures reproduced from arXiv: 2505.24634 by the authors.

Figure 1
Figure 1. Different partition mechanisms for voxel-based methods. From left to right: Cartesian Partition [10], [11], [4], [12], Polar/Cylindrical Partition [1], [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. (1) The framework of our method. We first obtain the voxel-wise feature using the proposed non-uniform cylindrical partition method. Then, the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Label encoding error of different partition methods. From left to right, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: The variation of points per cell with distance. Our method achieves [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: (a) The Non-uniform Multi-scale Aggregation. (b) The inconsistencies [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 7
Figure 7. Figure 7: We can observe that our method achieves superior [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 7
Figure 7. Figure 7: Comparison between the results of our method with and w/o non-uniform cylindrical partition(NUC) on SemanticKITTI validation set. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: The encoding errors with and w/o non-uniform partition on Se [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 10
Figure 10. Figure 10: The receptive field of 3×3 convolution in different partition methods. IX. ANALYSIS OF NON-UNIFORM CYLINDERICAL PARTITION A. Analysis of Performance In this section, we analyze why the proposed method outperforms the uniform partition methods and how point distributio…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

64 extracted references · 63 canonical work pages

  1. [1]

    Cylindrical and asymmetrical 3d convolution networks for lidar segmentation,

    X. Zhu, H. Zhou, T. Wang, F. Hong, Y . Ma, W. Li, H. Li, and D. Lin, “Cylindrical and asymmetrical 3d convolution networks for lidar segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021

  2. [2]

    (af)-s3net: Attentive feature fusion with adaptie feature selection for sparse semantic segmen- tation network,

    R. Cheng, R. Razani, E. Taghavi, E. Li, and B. Liu, “(af)-s3net: Attentive feature fusion with adaptie feature selection for sparse semantic segmen- tation network,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021

  3. [3]

    Polarnet: An improved grid representation for online lidar point clouds semantic segmentation,

    Y . Zhang, Z. Zhou, P. David, X. Yue, Z. Xi, B. Gong, and H. Foroosh, “Polarnet: An improved grid representation for online lidar point clouds semantic segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  4. [4]

    V oxelnet: End-to-end learning for point cloud based 3d object detection,

    Y . Zhou and O. Tuzel, “V oxelnet: End-to-end learning for point cloud based 3d object detection,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

  5. [5]

    Octnet: Learning deep 3d representation at high resolutions,

    G. Riegler, A. O. Ulusoy, and A. Geiger, “Octnet: Learning deep 3d representation at high resolutions,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017

  6. [6]

    V oxnet: a 3d convolutional neural network for real-time object recognition,

    D. Maturana and S. Scherer, “V oxnet: a 3d convolutional neural network for real-time object recognition,” inProceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2015

  7. [7]

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

    C. R. Qi, H. Su, K. Mo, and L. J. Guibas, “Pointnet: Deep learning on point sets for 3d classification and segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017

  8. [8]

    Pointnet++: Deep hierarchical feature learning on point sets in a metric space,

    C. R. Qi, L. Yi, H. Su, and L. J. Guibas, “Pointnet++: Deep hierarchical feature learning on point sets in a metric space,” inProceedings of the Conference and Workshop on Neural Information Processing Systems, 2017

Show all 64 references
  1. [9]

    3d semantic segmen- tation with submanifold sparse convolution networks,

    B. Graham, M. Engelcke, and L. van der Maaten, “3d semantic segmen- tation with submanifold sparse convolution networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

  2. [10]

    Search- ing efficient 3d architectures with sparse point-voxel convolution,

    H. Tang, Z. Liu, S. Zhao, Y . Lin, J. Lin, H. Wang, and S. Han, “Search- ing efficient 3d architectures with sparse point-voxel convolution,” in Proceedings of the European Conference on Computer Vision, 2020

  3. [11]

    2dpass: 2d priors assisted semantic segmentation on lidar point clouds,

    X. Yan, J. Gao, C. Zheng, C. Zheng, R. Zhang, S. Cui, and Z. Li, “2dpass: 2d priors assisted semantic segmentation on lidar point clouds,” inProceedings of the European Conference on Computer Vision, 2022

  4. [12]

    Link: Linear kernel for lidar-based 3d perception,

    T. Lu, X. Ding, H. Liu, G. Wu, and L. Wang, “Link: Linear kernel for lidar-based 3d perception,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2023

  5. [13]

    Spherical transformer for lidar-based 3d recognition,

    X. Lai, Y . Chen, F. Lu, J. Liu, and J. Jia, “Spherical transformer for lidar-based 3d recognition,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2023

  6. [14]

    Rangenet++: Fast and accurate lidar semantic segmentation,

    A. Milioto, I. Vizzo, J. Behley, and C. Stachniss, “Rangenet++: Fast and accurate lidar semantic segmentation,” inProceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2019

  7. [15]

    A deep and efficient range-point-voxel fusion network for lidar,

    J. Xu, R. Zhang, J. Dou, Y . Zhu, J. Sun, and S. Pu, “A deep and efficient range-point-voxel fusion network for lidar,” inProceedings of the IEEE International Conference on Computer Vision, 2021

  8. [16]

    Point-to-voxel knowledge distillation for lidar semantic segmentation,

    Y . Hou, X. Zhu, Y . Ma, C. C. Loy, and Y . Li, “Point-to-voxel knowledge distillation for lidar semantic segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2022

  9. [17]

    Drinet: A dual-representation iterative learning network for point cloud segmentation,

    M. Ye, S. Xu, T. Cao, and Q. Chen, “Drinet: A dual-representation iterative learning network for point cloud segmentation,” inProceedings of the IEEE International Conference on Computer Vision, 2021

  10. [18]

    Randla-net: Efficient semantic segmentation of large- scale point clouds,

    Q. Hu, B. Yang, L. Xie, S. Rosa, Y . Guo, Z. Wang, N. Trigoni, and A. Markham, “Randla-net: Efficient semantic segmentation of large- scale point clouds,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  11. [19]

    Real-time scene-aware lidar point cloud compression using semantic prior representation,

    L. Zhao, K.-K. Ma, Z. Liu, Q. Yin, and J. Chen, “Real-time scene-aware lidar point cloud compression using semantic prior representation,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 8, pp. 5623–5637, 2022

  12. [20]

    Towards robust lidar- camera fusion in bev space via mutual deformable attention and temporal aggregation,

    J. Wang, F. Li, Y . An, X. Zhang, and H. Sun, “Towards robust lidar- camera fusion in bev space via mutual deformable attention and temporal aggregation,”IEEE Transactions on Circuits and Systems for Video Technology, 2024

  13. [21]

    A multi-phase camera-lidar fusion network for 3d semantic segmentation with weak supervision,

    X. Chang, H. Pan, W. Sun, and H. Gao, “A multi-phase camera-lidar fusion network for 3d semantic segmentation with weak supervision,” IEEE Transactions on Circuits and Systems for Video Technology, 2023

  14. [22]

    Temporal-channel transformer for 3d lidar-based video object detection for autonomous driving,

    Z. Yuan, X. Song, L. Bai, Z. Wang, and W. Ouyang, “Temporal-channel transformer for 3d lidar-based video object detection for autonomous driving,”IEEE Transactions on Circuits and Systems for Video Technol- ogy, vol. 32, no. 4, pp. 2068–2078, 2021

  15. [23]

    Ffnet: Frequency fusion network for semantic scene completion,

    X. Wang, D. Lin, and L. Wan, “Ffnet: Frequency fusion network for semantic scene completion,” inProceedings of the AAAI conference on artificial intelligence, 2022

  16. [24]

    Squeezesegv3: Spatially-adaptive convolution for effi- cient point-cloud segmentation,

    C. Xu, B. Wu, Z. Wang, W. Zhan, P. Vajda, K. Keutzer, and M. Tomizuka, “Squeezesegv3: Spatially-adaptive convolution for effi- cient point-cloud segmentation,” inProceedings of the European Con- ference on Computer Vision, 2020

  17. [25]

    Rangevit: Towards vision transformers for 3d semantic segmentation in autonomous driving,

    A. Ando, S. Gidaris, A. Bursuc, G. Puy, A. Boulch, and R. Marlet, “Rangevit: Towards vision transformers for 3d semantic segmentation in autonomous driving,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2023

  18. [26]

    Adaint: Learning adaptive intervals for 3d lookup tables on real-time image enhancement,

    C. Yang, M. Jin, X. Jia, Y . Xu, and Y . Chen, “Adaint: Learning adaptive intervals for 3d lookup tables on real-time image enhancement,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2022

  19. [27]

    Deformable kernels: Adapting effective receptive fields for object deformation,

    H. Gao, X. Zhu, S. Lin, and J. Dai, “Deformable kernels: Adapting effective receptive fields for object deformation,” inProceedings of the International Conference on Learning Representations, 2020

  20. [28]

    Pointrend: image segmenta- tion as rendering,

    A. Kirillov, Y . Wu, K. He, and R. Girshick, “Pointrend: image segmenta- tion as rendering,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  21. [29]

    Hvnet: Hybrid voxel network for lidar based 3d object detection,

    M. Ye, S. Xu, and T. Cao, “Hvnet: Hybrid voxel network for lidar based 3d object detection,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  22. [30]

    Pointaugment: an auto- augmentation framework for point cloud classification,

    R. Li, X. Li, P.-A. Heng, and C.-W. Fu, “Pointaugment: an auto- augmentation framework for point cloud classification,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  23. [31]

    Panoptic-polarnet: Proposal-free lidar point cloud panoptic segmentation,

    Z. Zhou, Y . Zhang, and H. Foroosh, “Panoptic-polarnet: Proposal-free lidar point cloud panoptic segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021

  24. [32]

    Point-voxel cnn for efficient 3d deep learning,

    Z. Liu, H. Tang, Y . Lin, and S. Han, “Point-voxel cnn for efficient 3d deep learning,” inProceedings of the Conference and Workshop on Neural Information Processing Systems, 2019

  25. [33]

    The lovasz-softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks,

    M. Berman, A. R. Triki, and M. B. Blaschko, “The lovasz-softmax loss: A tractable surrogate for the optimization of the intersection-over-union measure in neural networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

  26. [34]

    Semantickitti: A dataset for semantic scene understanding of lidar sequences,

    J. Behley, M. Garbade, A. Milioto, J. Quenzel, S. Behnke, C. Stachniss, and J. Gall, “Semantickitti: A dataset for semantic scene understanding of lidar sequences,” inProceedings of the IEEE International Conference on Computer Vision, 2019

  27. [35]

    nuscenes: A multimodal dataset for autonomous driving,

    H. Caesar, V . Bankiti, A. H. Lang, S. V ora, V . E. Liong, Q. Xu, A. Kr- ishnan, Y . Pan, G. Baldan, and O. Beijbom, “nuscenes: A multimodal dataset for autonomous driving,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  28. [36]

    Efficient point cloud segmentation with geometry-aware sparse networks,

    M. Ye, R. Wan, S. Xu, T. Cao, and Q. Chen, “Efficient point cloud segmentation with geometry-aware sparse networks,” inProceedings of the European Conference on Computer Vision, 2022

  29. [37]

    Using a waffle iron for automotive point cloud semantic segmentation,

    G. Puy, A. Boulch, and R. Marlet, “Using a waffle iron for automotive point cloud semantic segmentation,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023

  30. [38]

    Rethinking range view representation for lidar segmentation,

    L. Kong, Y . Liu, R. Chen, Y . Ma, X. Zhu, Y . Li, Y . Hou, Y . Qiao, and Z. Liu, “Rethinking range view representation for lidar segmentation,” IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY , VOL., NO. 14 inProceedings of the IEEE/CVF International Conference...

  31. [39]

    Sfpnet: Sparse focal point network for semantic segmentation on general lidar point clouds

    Y . Wang, W. Zhao, C. Cao, T. Deng, J. Wang, and W. Chen., “Sfpnet: Sparse focal point network for semantic segmentation on general lidar point clouds.” inProceedings of the European Conference on Computer Vision, 2024

  32. [40]

    Sparse single sweep lidar point cloud segmentation via learning contextual shape priors from scene completion,

    X. Yan, J. Gao, J. Li, R. Zhang, Z. Li, R. Huang, and S. Cui, “Sparse single sweep lidar point cloud segmentation via learning contextual shape priors from scene completion,” inAAAI, 2021

  33. [41]

    Amvnet: Assertion-based multi-view fusion network for lidar semantic segmentation,

    V . E. Liong, T. N. T. Nguyen, S. Widjaja, D. Sharma, and Z. J. Chong, “Amvnet: Assertion-based multi-view fusion network for lidar semantic segmentation,” inarXiv preprint arXiv:2012.04934, 2021

  34. [42]

    Perception-aware multi-sensor fusion for 3d lidar semantic segmentation,

    Z. Zhuang, R. Li, K. Jia, Q. Wang, Y . Li, and M. Tan, “Perception-aware multi-sensor fusion for 3d lidar semantic segmentation,” inProceedings of the IEEE International Conference on Computer Vision, 2021

  35. [43]

    Learning 3d semantic segmen- tation with only 2d image supervision,

    K. Genova, X. Yin, A. Kundu, C. Pantofaru, F. Cole, A. Sud, B. Brew- ington, B. Shucker, and T. Funkhouser, “Learning 3d semantic segmen- tation with only 2d image supervision,” in3DV, 2021

  36. [44]

    Lidarmultinet: Towards a unified multi-task network for lidar percep- tion,

    D. Ye, Z. Zhou, W. Chen, Y . Xie, Y . Wang, P. Wang, and H. Foroosh, “Lidarmultinet: Towards a unified multi-task network for lidar percep- tion,” inarXiv preprint arXiv:2206.11428, 2022

  37. [45]

    Tanget convolu- tions for dense prediction in 3d,

    M. Tatarchenko, J. Park, V . Koltun, and Q.-Y . Zhou, “Tanget convolu- tions for dense prediction in 3d,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

  38. [46]

    Spsequencenet: Semantic segmentation network on 4d point clouds,

    H. Shi, G. Lin, H. Wang, T.-Y . Hung, and Z. Wang, “Spsequencenet: Semantic segmentation network on 4d point clouds,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020

  39. [47]

    Kpconv: Flexible and deformable convolution for point cloud,

    H. Thomas, C. R. Qi, J.-E. Deschaud, B. Marcotegui, and F. G. L. J. Guibas, “Kpconv: Flexible and deformable convolution for point cloud,” inProceedings of the IEEE International Conference on Computer Vision, 2019

  40. [48]

    Pv- rcnn: Point-voxel feature set abstraction for 3d object detection,

    S. Shi, C. Guo, L. Jiang, Z. Wang, J. Shi, X. Wang, and H. Li, “Pv- rcnn: Point-voxel feature set abstraction for 3d object detection,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  41. [49]

    Pointgroup: Dual-set point grouping for 3d instance segmentation,

    L. Jiang, H. Zhao, S. Shi, S. Liu, C.-W. Fu, and J. Jia, “Pointgroup: Dual-set point grouping for 3d instance segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2020

  42. [50]

    Lidar panoptic segmentation for autonomous driving,

    A. Milioto, J. Behley, C. McCool, and C. Stachniss, “Lidar panoptic segmentation for autonomous driving,” inProceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2020

  43. [51]

    Scribble-supervised lidar semantic segmentation,

    O. Unal, D. Dai, and L. V . Gool, “Scribble-supervised lidar semantic segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2022

  44. [52]

    4d spatio-temporal convnets: Minkowski convolutional neural networks,

    C. Choy, J. Gwak, and S. Savarese, “4d spatio-temporal convnets: Minkowski convolutional neural networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019

  45. [53]

    Squeezeseg: Convolutional neural nets with recurrent crf for real-time road-object segmentation from 3d lidar point cloud,

    B. Wu, A. Wan, X. Yue, and K. Keutzer, “Squeezeseg: Convolutional neural nets with recurrent crf for real-time road-object segmentation from 3d lidar point cloud,” in2018 IEEE International Conference on Robotics and Automation (ICRA), 2018

  46. [54]

    Squeezesegv2: Improved model structure and unsupervised domain adaptation for road- object segmentation from a lidar point cloud,

    B. Wu, X. Zhou, S. Zhao, X. Yue, and K. Keutzer, “Squeezesegv2: Improved model structure and unsupervised domain adaptation for road- object segmentation from a lidar point cloud,” in2019 international conference on robotics and automation (ICRA), 2019

  47. [55]

    Salsanext: Fast, uncertainty- aware semantic segmentation of lidar point clouds,

    T. Cortinhal, G. Tzelepis, and E. E. Aksoy, “Salsanext: Fast, uncertainty- aware semantic segmentation of lidar point clouds,” inInternational Symposium on Visual Computing, 2020

  48. [56]

    Fidnet: Lidar point cloud semantic segmentation with fully interpolation decoding,

    Y . Zhao, L. Bai, and X. Huang, “Fidnet: Lidar point cloud semantic segmentation with fully interpolation decoding,” in2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021

  49. [57]

    Cenet: Toward concise and efficient lidar semantic segmentation for autonomous driving,

    H. Cheng, X. Han, and G. Xiao, “Cenet: Toward concise and efficient lidar semantic segmentation for autonomous driving,” in2022 IEEE international conference on multimedia and expo (ICME), 2022

  50. [58]

    Pids: Joint point interaction-dimension search for 3d point cloud,

    T. Zhang, M. Ma, F. Yan, H. Li, and Y . Chen, “Pids: Joint point interaction-dimension search for 3d point cloud,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023

  51. [59]

    Cpgnet: Cascade point-grid fusion network for real-time lidar semantic segmentation,

    X. Li, G. Zhang, H. Pan, and Z. Wang, “Cpgnet: Cascade point-grid fusion network for real-time lidar semantic segmentation,” in2022 International Conference on Robotics and Automation (ICRA), 2022

  52. [60]

    Gfnet: Geometric flow network for 3d point cloud semantic segmentation,

    H. Qiu, B. Yu, and D. Tao, “Gfnet: Geometric flow network for 3d point cloud semantic segmentation,”Transactions on Machine Learning Research, 2022

  53. [61]

    Panoptic segmentation,

    A. Kirillov, K. He, R. Girshick, C. Rother, and P. Dollar, “Panoptic segmentation,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019

  54. [62]

    Embracing single stride 3d object detector with sparse transformer,

    L. Fan, Z. Pang, T. Zhang, Y .-X. Wang, H. Zhao, F. Wang, N. Wang, and Z. Zhang, “Embracing single stride 3d object detector with sparse transformer,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2022

  55. [63]

    3d semantic segmentation with submanifold sparse convolutional networks,

    Y . Yan, Y . Mao, and B. Li, “3d semantic segmentation with submanifold sparse convolutional networks,” inSensors, 2018

  56. [64]

    3d semantic segmen- tation with submanifold sparse convolutional networks,

    B. Graham, M. Engelcke, and L. van der Maaten, “3d semantic segmen- tation with submanifold sparse convolutional networks,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018

Pith tools

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