Pith. sign in

REVIEW 3 major objections 6 minor 64 references

PointODE: Lightweight Point Cloud Learning with Neural Ordinary Differential Equations on Edge

T0 review · 3 major / 6 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 A shared ODE block compresses a point cloud classifier to 0.58M parameters while keeping accuracy and speeding up FPGA inference 3.7x.

desk verdict A credible, well-scoped embedded point cloud paper whose headline FPGA speedups should be read as 'vs PyTorch on an ARM CPU,' not as an absolute win. read the letter →

arxiv 2506.00438 v1 pith:DZXOXJ3I submitted 2025-05-31 cs.LG cs.AR

Keisuke Sugiura , Mizuki Yasuda , Hiroki Matsutani This is my paper · ORCID
classification cs.LGcs.AR
keywords pointcloudclassificationNeuralODEparametersharingresidualMLPFPGAacceleratorpoint-wisenormalizationedgeinferenceModelNet40
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to show that a point cloud classifier can be made small enough for an embedded FPGA without sacrificing accuracy by treating residual MLP blocks as one shared ODE block and by normalizing each point's feature vector independently. The resulting PointODE-Elite has 0.58M trainable parameters, keeps 93.4% overall accuracy on ModelNet40 and 84.2% on ScanObjectNN, and its FPGA implementation runs feature extraction 4.9x faster and full inference 3.7x faster than the ARM Cortex-A53 CPU baseline, with 3.5x better energy efficiency. If correct, this gives edge devices a practical way to run point cloud classification with a much better accuracy-versus-cost trade-off than current models. The central mechanism is parameter sharing across repeated iterations of a single block plus per-point normalization, which removes the need to wait for global statistics and enables a four-stage pipeline.

What carries the argument

The central object is the ODEPBlock, an ODE-based residual point block: each forward iteration concatenates a time variable to the feature, applies FC-BN-ReLU, concatenates time again, applies FC-BN, and adds the result to the input, matching the Euler update $h(t_j)=h(t_{j-1})+h f(h(t_{j-1}),t_{j-1},\theta)$. Because the same parameters $\theta$ are reused for $C$ iterations, the block behaves like $C$ residual blocks sharing weights, which is where the parameter reduction comes from. The other load-bearing mechanism is point-wise normalization, Eqs. (5)-(6): for each neighboring feature vector the mean and standard deviation are computed over feature dimensions only, replacing PointMLP's global $\mu$ and $\sigma$. This makes every point's normalization independent, removes the synchronization barrier before each MLP stage, and is what the four-stage FPGA pipeline exploits.

What would settle it

Compute PointODE-Elite's accuracy on a deliberately non-uniformly sampled or noisy variant of ModelNet40 and compare it with PointMLP under the same training protocol: if the gap exceeds the reported 0.1-1.1% range, or if reintroducing global statistics removes the FPGA's 4.9x feature-extraction speedup, the central premise fails.

Watch

Extended reading notes

Core claim

The paper claims that replacing PointMLP's four separate residual point blocks per stage with repeated forward iterations of a single ODE-based residual point block (ODEPBlock), after reordering blocks so they are consecutive, compresses the model by 23x overall while keeping accuracy almost unchanged: PointODE-Elite reaches 93.4% OA on ModelNet40 and 84.2% on ScanObjectNN, within 0.1-1.1% of PointMLP. It further claims that swapping PointMLP's global geometric affine transform for point-wise normalization, which computes mean and standard deviation per feature vector rather than across all points, both improves accuracy on the real-world ScanObjectNN benchmark and lets each stage process sampled points independently. That independence is what allows the FPGA accelerator to pipeline four points concurrently and to keep the entire model and intermediate buffers on-chip. The paper reports measured 4.9x feature extraction speedup, 3.7x full inference speedup, and 3.5x energy efficiency gain on a Xilinx ZCU104 board while preserving accuracy under 24-bit fixed-point arithmetic.

Load-bearing premise

The load-bearing premise is that replacing PointMLP's global geometric affine transform with per-point normalization is at least as accurate and makes each point independently processable; this premise is supported only by a ScanObjectNN ablation showing a 1.1-1.5% accuracy gain, with no formal guarantee for non-uniform or noisy point clouds.

Editorial extensions

If this is right

  • PointODE-Elite keeps classification accuracy within 0.1-1.1% of PointMLP while using 23x fewer parameters and 25.5x fewer FLOPs, and it matches or beats PointMLP-Elite with 1.25x fewer parameters.
  • On the ZCU104 FPGA, feature extraction is 4.9x faster than the ARM Cortex-A53 CPU, full inference is 3.7x faster, and energy efficiency is 3.5x better, with gains coming from on-chip storage of all parameters and four-point pipelining.
  • Increasing the number of ODE iterations $C$ from 1 to 8 raises ScanObjectNN accuracy from 83.6% to 84.7% at linearly growing cost, while $C\ge10$ causes accuracy to drop, consistent with accumulated Euler solver error.
  • The full model and its intermediate buffers fit in on-chip memory (95% of URAM and 69% of BRAM used on ZCU104), so most off-chip data transfers are eliminated.
  • A 24-bit fixed-point FPGA implementation preserves accuracy (93.7% OA on ModelNet40), suggesting that more aggressive quantization could be applied.

Reading between the lines

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

  • Our inference: point-wise normalization could improve robustness to non-uniform sampling and partial occlusion, since each point adapts its own scaling factor; this is testable by evaluating PointODE-Elite on increasingly noisy versions of ModelNet40 or on real LiDAR scans.
  • Our inference: because the ODE iterations share one block and each point is processed independently, the architecture could be adapted to streaming point cloud input, processing points as they arrive rather than after full-cloud accumulation; the paper does not explore this.
  • Our inference: higher-order ODE solvers or adaptive step sizes would likely extend the useful range of $C$ beyond 8 and close the remaining accuracy gap to PointMLP, since the paper attributes the drop at $C\ge10$ to numerical error.
  • Our inference: the same parameter-sharing recipe may compress other residual point cloud models, including point cloud transformers, wherever consecutive residual blocks dominate parameter count; the paper only demonstrates it for PointMLP-style MLPs.
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 manuscript proposes PointODE, a point cloud classification architecture obtained by applying Neural ODE parameter sharing to PointMLP's residual blocks, together with a point-wise normalization variant and a lightweight 0.58M-parameter version, PointODE-Elite. The authors report competitive classification accuracy on ModelNet40 and ScanObjectNN (Table 1), an ablation isolating reordering, normalization, and dimension reduction (Table 2), close parity between FPGA fixed-point and PyTorch accuracy (Table 3), and an FPGA implementation on ZCU104 that achieves 4.9x feature-extraction speedup, 3.7x end-to-end speedup, and 3.5x energy-efficiency gain versus an ARM Cortex-A53 CPU baseline (Sec. 6.3-6.4). The central claim is that the combination of parameter sharing and point-wise normalization yields a practical accuracy-versus-inference-cost trade-off for edge point cloud classification.

Significance. If the hardware results hold, this is a useful systems contribution: it demonstrates that Neural ODE-style parameter sharing can compress a point cloud architecture substantially without a large accuracy penalty, and that point-wise normalization enables a four-stage FPGA pipeline that processes multiple points independently. The paper has notable strengths: it evaluates on standard datasets with external baselines, provides an ablation that isolates each architectural change, reports FPGA accuracy that closely matches the PyTorch model despite fixed-point quantization, and gives resource utilization numbers supporting the on-chip storage claim. The main risk is not internal inconsistency in the architecture but the validity of the hardware comparison: the speedup and energy gains are measured against an underspecified PyTorch-on-ARM baseline, so the cost-side numbers need additional validation before the trade-off claim is fully established.

major comments (3)
  1. [Sec. 6.3, Figs. 12-13] The headline speedups (4.9x feature extraction, 3.7x end-to-end, 3.5x energy) are computed against an ARM Cortex-A53 baseline identified only as 'PyTorch-based' (Sec. 6.4). The paper does not report whether the CPU code is single- or multi-threaded, whether NEON/SIMD is used, whether Python interpreter overhead contributes, or what compiler/optimization flags are used. A 337.5 ms feature-extraction time for a 0.58M-parameter MLP on 1024 points is far above what an optimized C/C++ fixed-point implementation would achieve on a 1.2 GHz quad-core A53. Because the paper's central claim is an accuracy-versus-inference-cost trade-off, the CPU baseline must be replaced or supplemented by a carefully optimized CPU implementation (for example, the same fixed-point kernel compiled in C with single-core and multi-core variants) and the comparison repeated.
  2. [Tables 1-3, Sec. 6.1.3] All accuracy numbers are single runs, and the ODE integration interval tb and iteration count C are chosen per dataset from the validation accuracy of the same datasets. The ablation in Table 2 reports differences of 1.1-1.5% for point-wise normalization; without multiple seeds or error bars, it is not possible to tell whether these differences are significant. At minimum, the authors should report mean and standard deviation over at least three seeds for the main models and the ablation, and either fix tb and C before evaluation or describe a hold-out procedure.
  3. [Sec. 4.2 and footnote 1] The text says 'each stage has only one ODEPBlock instead of two' but the footnote states 'The first stage keeps two separate ODEPBlocks for accuracy'. This contradiction affects the parameter-count claims and the description of how many blocks are actually shared by the ODE. Please clarify how many ODEPBlocks exist per stage in PointODE and PointODE-Elite and provide a breakdown of parameters per component; otherwise the 23.02x/25.51x reductions reported in Fig. 1 and Sec. 4.4 cannot be fully audited from the text.
minor comments (6)
  1. [Sec. 3.1] The word 'buliding' should be 'building'.
  2. [Eq. (6)] The rendering of the expression for tilde-sigma appears to have a stray subscript after the norm term; please ensure the point-wise mean and deviation are defined without ambiguity.
  3. [Table 2] The row labels are ambiguous; each ablation row should be labeled explicitly (e.g., '+Reorder', '+Point-wise Norm', '+Reduced Dims') so the reader can reconstruct the incremental effects.
  4. [Sec. 4.4] The phrase '1.25x fewer parameters and 1.83x fewer FLOPs' is imprecise; consider '1.25x lower parameter count and 1.83x lower FLOPs' or give the percentage reductions.
  5. [Sec. 6.4] The power sentence 'PyTorch- and FPGA-based PointODE-Elite consume 0.43W and 0.45W of power (from 11.35W to 11.78W and 11.80W)' is hard to parse; state idle power and incremental power separately for each implementation.
  6. [Sec. 6.1.3] For the results in Table 1 and Fig. 11, the paper should state which tb value is used in each experiment, because the final time is selected from {0.1, 0.2, 0.3}.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the paper is an empirical architecture and hardware evaluation against external benchmarks and baselines.

full rationale

The paper's load-bearing claims are experimental: classification accuracy of PointODE-Elite on ModelNet40 and ScanObjectNN is compared with external baselines (PointMLP, PointNeXt, RepSurf, etc.), and the FPGA speedups are measured against an ARM Cortex-A53 CPU implementation. The ODE parameter-sharing construction (Secs. 4.1-4.4) directly applies the long-established ResNet/Neural ODE equivalence (Eqs. 1-2, citing Chen et al. [19]) rather than deriving a result from assumptions that include its own conclusion. The point-wise normalization is validated by the controlled ablation in Table 2, not by an equation that reduces to the claim. Self-citations [56,57] appear only in related work describing prior FPGA Neural ODE designs; they are not used to justify the architecture or forbid alternatives. The hyperparameter choices tb in {0.1,0.2,0.3} and iteration count C (Sec. 6.1.3, Fig. 11) are model-selection decisions on the evaluation datasets, which is a reporting-quality concern rather than circularity, since the reported numbers are still measured results on external benchmarks. No step of the derivation chain is equivalent to its inputs by construction.

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

Headline accuracy and hardware performance depend on user-selected tb, C, the 24-bit fixed-point format, and the hand-picked architecture widths. The accuracy claims are empirical and not derived. No new physical entities are introduced. The main modeling premises are that ODE weight-sharing preserves capacity and that per-point normalization is at least as accurate as global affine normalization; both are supported only by ablations.

free parameters (4)
  • ODE integration interval final time tb = 0.1, 0.2, or 0.3, chosen per model and dataset
    Sec. 6.1.3 selects tb from {0.1,0.2,0.3} to achieve the best accuracy, and the reported accuracy numbers depend on this choice.
  • ODE iteration count C = 1 to 10; C=4 for the reported FPGA results
    Fig. 11 shows accuracy improving with C up to 8 and dropping for C>=10; C=4 is used in the timing and power evaluation (Sec. 6.3), so the reported speed and accuracy points are coupled to this choice.
  • Fixed-point format = 24-bit, 8-bit integer and 16-bit fractional
    Sec. 6.1.1 chooses this representation by hand to keep FPGA accuracy close to PyTorch; it is a free design parameter of the hardware accuracy claim.
  • PointODE-Elite widths and group size = F1-F4=(64,128,256,256), F0=32, K=12, bottleneck F'_s=Fs/4
    Architecture size choices from Sec. 4.4 that define the 0.58M parameter count; they are chosen by hand following PointMLP-Elite and affect both accuracy and on-chip FPGA fit.
assumptions (4)
  • standard math Euler discretization (Eq. 2) is a sufficient ODE solver for the compressed residual network.
    The ODEPBlock forward pass and the FPGA design both implement forward Euler; no convergence or stability analysis is given, and Fig. 11 shows accuracy collapse for large C.
  • domain assumption ExNODE's permutation-invariance theorem applies to ODEPBlock because all operations act per point.
    Sec. 4.4 cites [52] for the invariant feature property and checks each layer separately; no independent proof is given in this paper.
  • domain assumption Per-point normalization (Eqs. 5-6) preserves or improves the accuracy of PointMLP's global affine normalization (Eqs. 3-4).
    This premise is load-bearing for the pipelined FPGA design; its support is only the ScanObjectNN ablation in Table 2, with no theoretical justification.
  • domain assumption The training recipe inherited from PointMLP (learning rates, cosine schedule, augmentations) transfers to weight-tied ODE blocks.
    Sec. 6.1.3 uses standard PointMLP-style training without analyzing optimization of shared parameters; convergence is assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PointODE: Lightweight Point Cloud Learning with Neural Ordinary Differential Equations on Edge." pith.science (2026). https://pith.science/paper/DZXOXJ3I

@misc{pith2026250600438,
  author       = {Pith},
  title        = {Pith review of: PointODE: Lightweight Point Cloud Learning with Neural Ordinary Differential Equations on Edge},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DZXOXJ3I}},
  note         = {Machine review of arXiv:2506.00438}
}
read the original abstract

Embedded edge devices are often used as a computing platform to run real-world point cloud applications, but recent deep learning-based methods may not fit on such devices due to limited resources. In this paper, we aim to fill this gap by introducing PointODE, a parameter-efficient ResNet-like architecture for point cloud feature extraction based on a stack of MLP blocks with residual connections. We leverage Neural ODE (Ordinary Differential Equation), a continuous-depth version of ResNet originally developed for modeling the dynamics of continuous-time systems, to compress PointODE by reusing the same parameters across MLP blocks. The point-wise normalization is proposed for PointODE to handle the non-uniform distribution of feature points. We introduce PointODE-Elite as a lightweight version with 0.58M trainable parameters and design its dedicated accelerator for embedded FPGAs. The accelerator consists of a four-stage pipeline to parallelize the feature extraction for multiple points and stores the entire parameters on-chip to eliminate most of the off-chip data transfers. Compared to the ARM Cortex-A53 CPU, the accelerator implemented on a Xilinx ZCU104 board speeds up the feature extraction by 4.9x, leading to 3.7x faster inference and 3.5x better energy-efficiency. Despite the simple architecture, PointODE-Elite shows competitive accuracy to the state-of-the-art models on both synthetic and real-world classification datasets, greatly improving the trade-off between accuracy and inference cost.

Figures

Figures reproduced from arXiv: 2506.00438 by the authors.

Figure 1
Figure 1. Number of parameters and FLOPs of the proposed PointODE-Elite and the baseline PointMLP. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Replacing the forward pass of C ResBlocks with C forward iterations of the single ODE-based building block. high-dimensional features for a subset of points, which are fed to the network for a specific task (e.g., classification and segmentation). PointMLP consists of an embedding block to extract F0-dim local features {f 0 1 , . . . ,f 0 N } for each point, which is followed by a stack of four stages to hierarchica… view at source ↗
Figure 3
Figure 3. Architecture of PointMLP. Each stage s ∈ [1, 4] produces Fs-dim features F s = {f s i } for Ns sampled points P s = {p s i }. PointMLP first samples Ns points (i.e., group centroids) from P s−1 and finds their K-nearest neighbors (NNs). For each group centroid p s−1 j ∈ Ps−1 , the geometric affine module applies an affine transformation to the features {f s−1 j,k } of its neighbors {p s−1 j,k }. The transformed feat… view at source ↗
Figures from the paper (10 more)
Figure 4
Figure 4. Figure 4: ResPBlock and ODEPBlock. 4.2 Residual Block Reordering To improve the effectiveness of using ODE blocks, we propose to first reorder the blocks in each stage of PointMLP before employing Neural ODE. As shown in [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Residual block reordering (left: PointMLP, center: reordered layers, right: PointODE(-Elite)). [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Architecture of PointODE and PointODE-Elite. [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: Block diagram of modules for the point-wise normalization, MLP block (FC-BN-ReLU), and max-pooling. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Block diagram of ODEPBlock. normalization (Eqs. 5–6) instead of the geometric affine module (Eqs. 3–4), each sampled point can be processed independently. Unlike PointMLP, each stage can start running the subsequent MLP and ODEPBlock without waiting for the computation…
Figure 9
Figure 9. Figure 9: Block diagram of the stage. 5.4 Implementation of PointODE-Elite [PITH_FULL_IMAGE:figures/full_fig_p008_9.png]
Figure 10
Figure 10. Figure 10: Overview of the FPGA implementation. in Sec. 5.1, each stage s needs precomputed indices of size (Ns, K) to subsample Ns points P s out of Ns−1 points P s−1 and find KNNs for each sampled point. The host performs farthest point sampling (FPS) and KNN search recursivel…
Figure 11
Figure 11. Figure 11: plots the accuracy of PointODE-Elite with respect to the number of ODE iterations C on ScanObjectNN. By increasing C from 1 to 8, the overall accuracy improves by 1.1% (83.6% to 84.7%) at the cost of linearly increasing computational cost in ODEPBlocks, which closes t…
Figure 12
Figure 12. Figure 12: Time for feature extraction with respect to the number of ODE iterations [PITH_FULL_IMAGE:figures/full_fig_p012_12.png]
Figure 13
Figure 13. Figure 13: Execution time breakdown (N = 1024, C = 4). the BRAM and URAM utilization, which in turn allows to assign more memory blocks to each buffer and increase the number of read/write ports. Considering that DSP blocks are underutilized, the design could further parallelize…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 61 canonical work pages

  1. [1]

    Robust Reconstruction of Indoor Scenes

    Sungjoon Choi, Qian-Yi Zhou, and Vladlen Koltun. Robust Reconstruction of Indoor Scenes. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5556–5565, June 2015. 12 PointODE: Lightweight Point Cloud Learning with Neural ODEs on Edge A PREPRINT

    work page 2015

  2. [2]

    Colored Point Cloud Registration Revisited

    Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. Colored Point Cloud Registration Revisited. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), pages 143–152, October 2017

    work page 2017

  3. [3]

    LOAM: Lidar Odometry and Mapping in Real-time

    Ji Zhang and Sanjiv Singh. LOAM: Lidar Odometry and Mapping in Real-time. In Proceedings of the Robotics: Science and Systems (RSS), pages 1–9, July 2014

    work page 2014

  4. [4]

    LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Map- ping on Variable Terrain

    Tixiao Shan and Brendan Englot. LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Map- ping on Variable Terrain. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 4758–4765, October 2018

    work page 2018

  5. [5]

    LIO-SAM: Tightly- coupled Lidar Intertial Odometry via Smoothing and Mapping

    Tixiao Shan, Brendan Englot, Drew Meyers, Wei Wang, Carlo Ratti, and Daniela Rus. LIO-SAM: Tightly- coupled Lidar Intertial Odometry via Smoothing and Mapping. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5135–5142, October 2020

    work page 2020

  6. [6]

    F-LOAM: Fast LiDAR Odometry and Mapping

    Han Wang, Chen Wang, Chun-Lin Chen, and Lihua Xie. F-LOAM: Fast LiDAR Odometry and Mapping. In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) , pages 4390– 4396, September 2021

    work page 2021

  7. [7]

    Tracking Objects as Points

    Xingyi Zhou, Vladlen Koltun, and Philipp Krähenbühl. Tracking Objects as Points. In Proceedings of the European Conference on Computer Vision (ECCV), pages 474–490, August 2020

    work page 2020

  8. [8]

    Center-Based 3D Object Detection and Tracking

    Tianwei Yin, Xingyi Zhou, and Philipp Krähenbühl. Center-Based 3D Object Detection and Tracking. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11784– 11793, June 2021

    work page 2021

Show all 64 references
  1. [9]

    Qi, Hao Su, Kaichun Mo, and Leonidas J

    Charles R. Qi, Hao Su, Kaichun Mo, and Leonidas J. Guibas. PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 652–660, July 2017

    2017

  2. [10]

    PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space

    Charles Ruizhongtai Qi, Li Yi, Hao Su, and Leonidas Guibas. PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space. In Proceedings of the Advances in Neural Information Processing Systems (NIPS), pages 5099–5108, December 2017

    2017

  3. [11]

    FPConv: Learning Local Flattening for Point Convolution

    Yiqun Lin, Zizheng Yan, Haibin Huang, Dong Du, Ligang Liu, Shuguang Cui, and Xiaoguang Han. FPConv: Learning Local Flattening for Point Convolution. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 4293–4302, June 2020

    2020

  4. [12]

    PAConv: Position Adaptive Convolution With Dynamic Kernel Assembling on Point Clouds

    Mutian Xu, Runyu Ding, Hengshuang Zhao, and Xiaojuan Qi. PAConv: Position Adaptive Convolution With Dynamic Kernel Assembling on Point Clouds. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 3173–3182, June 2021

    2021

  5. [13]

    Adaptive Graph Convo- lution for Point Cloud Analysis

    Haoran Zhou, Yidan Feng, Mingsheng Fang, Mingqiang Wei, Jing Qin, and Tong Lu. Adaptive Graph Convo- lution for Point Cloud Analysis. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 4965–4974, October 2021

    2021

  6. [14]

    Learning of 3D Graph Convolution Networks for Point Cloud Analysis

    Zhi-Hao Lin, Sheng-Yu Huang, and Yu-Chiang Frank Wang. Learning of 3D Graph Convolution Networks for Point Cloud Analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 44(8):4212– 4224, August 2022

    2022

  7. [15]

    Point-BERT: Pre-Training 3D Point Cloud Transformers With Masked Point Modeling

    Xumin Yu, Lulu Tang, Yongming Rao, Tiejun Huang, Jie Zhou, and Jiwen Lu. Point-BERT: Pre-Training 3D Point Cloud Transformers With Masked Point Modeling. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 19313–19322, June 2022

    2022

  8. [16]

    Dual Transformer for Point Cloud Analysis

    Xian-Feng Han, Yi-Fei Jin, Hui-Xian Cheng, and Guo-Qiang Xiao. Dual Transformer for Point Cloud Analysis. IEEE Transactions on Multimedia (TMM), 25(1):5638–5648, August 2022

    2022

  9. [17]

    Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework

    Xu Ma, Can Qin, Haoxuan You, Haoxi Ran, and Yun Fu. Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework. In Proceedings of the International Conference on Learning Representations (ICLR), pages 1–15, April 2022

    2022

  10. [18]

    Deep Residual Learning for Image Recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep Residual Learning for Image Recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770–778, June 2016

    2016

  11. [19]

    Ricky T. Q. Chen, Yulia Rubanova, Jesse Bettencourt, and David K. Duvenaud. Neural Ordinary Differential Equations. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS) , pages 6571– 6583, December 2018

    2018

  12. [20]

    Yulia Rubanova, Ricky T. Q. Chen, and David K. Duvenaud. Latent Ordinary Differential Equations for Irregularly-Sampled Time Series. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), pages 5320–5330, December 2019. 13 PointODE: Lightweight Poin...

    2019

  13. [21]

    Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks

    Zonghan Wu, Shirui Pan, Guodong Long, Jing Jiang, Xiaojun Chang, and Chengqi Zhang. Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD), pages...

    2020

  14. [22]

    Neural Controlled Differential Equations for Irregular Time Series

    Patrick Kidger, James Morrill, James Foster, and Terry Lyons. Neural Controlled Differential Equations for Irregular Time Series. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS) , pages 6696–6707, December 2020

    2020

  15. [23]

    Modeling Tra- jectories with Neural Ordinary Differential Equations

    Yuxuan Liang, Kun Ouyang, Hanshu Yan, Yiwei Wang, Zekun Tong, and Roger Zimmermann. Modeling Tra- jectories with Neural Ordinary Differential Equations. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI), pages 1498–1504, August 2021

    2021

  16. [24]

    Social ODE: Multi-agent Trajectory Forecasting with Neural Ordinary Differential Equations

    Song Wen, Hao Wang, and Dimitris Metaxas. Social ODE: Multi-agent Trajectory Forecasting with Neural Ordinary Differential Equations. In Proceedings of the European Conference on Computer Vision (ECCV) , pages 217–233, October 2022

    2022

  17. [25]

    Climate Modeling with Neural Diffusion Equations

    Jeehyun Hwang, Jeongwhan Choi, Hwangyong Choi, Kookjin Lee, Dongeun Lee, and Noseong Park. Climate Modeling with Neural Diffusion Equations. In Proceedings of the IEEE International Conference on Data Mining (ICDM), pages 230–239, December 2021

    2021

  18. [26]

    ClimODE: Climate and Weather Forecasting With Physics- informed Neural ODEs

    Yogesh Verma, Markus Heinonen, and Vikas Garg. ClimODE: Climate and Weather Forecasting With Physics- informed Neural ODEs. In Proceedings of the International Conference on Learning Representations (ICLR) , pages 1–23, May 2024

    2024

  19. [27]

    Guangsi Shi, Daokun Zhang, Ming Jin, Shirui Pan, and Philip S. Yu. Towards Complex Dynamic Physics System Simulation with Graph Neural Ordinary Equations. Neural Networks, 176(C):106341, August 2024

    2024

  20. [28]

    Vid-ODE: Continuous-Time Video Generation with Neural Ordinary Differential Equation

    Sunghyun Park, Kangyeol Kim, Junsoo Lee, Jaegul Choo, Joonseok Lee, Sookyung Kim, and Edward Choi. Vid-ODE: Continuous-Time Video Generation with Neural Ordinary Differential Equation. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 2412–2422, February 2021

    2021

  21. [29]

    TiV-ODE: A Neural ODE-based Approach for Controllable Video Generation From Text-Image Pairs

    Yucheng Xu, Nanbo Li, Arushi Goel, Zonghai Yao, Zijian Guo, and Hamidreza Kasaei. TiV-ODE: A Neural ODE-based Approach for Controllable Video Generation From Text-Image Pairs. InIEEE International Confer- ence on Robotics and Automation (ICRA), pages 14645–14652, May 2024

    2024

  22. [30]

    Jiahao, Jiancong Wang, Paul A

    Yifan Wu, Tom Z. Jiahao, Jiancong Wang, Paul A. Yushkevich, M. Ani Hsieh, and James C. Gee. NODEO: A Neural Ordinary Differential Equation Based Optimization Framework for Deformable Image Registration. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern R...

    2022

  23. [31]

    Point Convolutional Neural Networks by Extension Opera- tors

    Matan Atzmon, Haggai Maron, and Yaron Lipman. Point Convolutional Neural Networks by Extension Opera- tors. ACM Transactions on Graphics (TOG), 37(4):1–12, August 2018

    2018

  24. [32]

    SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters

    Yifan Xu, Tianqi Fan, Mingye Xu, Long Zeng, and Yu Qiao. SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters. In Proceedings of the European Conference on Computer Vision (ECCV), pages 87–102, September 2018

    2018

  25. [33]

    PointCNN: Convolution On X-Transformed Points

    Yangyan Li, Rui Bu, Mingchao Sun, Wei Wu, Xinhan Di, and Baoquan Chen. PointCNN: Convolution On X-Transformed Points. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS) , pages 820–830, December 2018

    2018

  26. [34]

    Relation-Shape Convolutional Neural Network for Point Cloud Analysis

    Yongcheng Liu, Bin Fan, Shiming Xiang, and Chunhong Pan. Relation-Shape Convolutional Neural Network for Point Cloud Analysis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8895–8904, June 2019

    2019

  27. [35]

    PointConv: Deep Convolutional Networks on 3D Point Clouds

    Wenxuan Wu, Zhongang Qi, and Li Fuxin. PointConv: Deep Convolutional Networks on 3D Point Clouds. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 9621– 9630, June 2019

    2019

  28. [36]

    Sarma, Michael M

    Yue Wang, Yongbin Sun, Ziwei Liu, Sanjay E. Sarma, Michael M. Bronstein, and Justin M. Solomon. Dynamic Graph CNN for Learning on Point Clouds. ACM Transactions on Graphics (TOG), 38(5):1–12, October 2019

    2019

  29. [37]

    Grid-GCN for Fast and Scalable Point Cloud Learning

    Qiangeng Xu, Xudong Sun, Cho-Ying Wu, Panqu Wang, and Ulrich Neumann. Grid-GCN for Fast and Scalable Point Cloud Learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 5661–5670, June 2020

    2020

  30. [38]

    Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds

    Huan Lei, Naveed Akhtar, and Ajmal Mian. Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 43(10):3664–3680, October 2021. 14 PointODE: Lightweight Point Cloud Learning with Neural ODEs ...

    2021

  31. [39]

    Martin, and Shi-Min Hu

    Meng-Hao Guo, Jun-Xiong Cai, Zheng-Ning Liu, Tai-Jiang Mu, Ralph R. Martin, and Shi-Min Hu. PCT: Point Cloud Transformer. Computational Visual Media, 7(2):187–199, June 2021

    2021

  32. [40]

    Torr, and Vladlen Koltun

    Hengshuang Zhao, Li Jiang, Jiaya Jia, Philip H.S. Torr, and Vladlen Koltun. Point Transformer. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 16259–16268, October 2021

    2021

  33. [41]

    Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis

    Tiange Xiang, Chaoyi Zhang, Yang Song, Jianhui Yu, and Weidong Cai. Walk in the Cloud: Learning Curves for Point Clouds Shape Analysis. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 915–924, October 2021

    2021

  34. [42]

    MeshNet: Mesh Neural Network for 3D Shape Representation

    Yutong Feng, Yifan Feng, Haoxuan You, Xibin Zhao, and Yue Gao. MeshNet: Mesh Neural Network for 3D Shape Representation. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 8279–8286, July 2019

    2019

  35. [43]

    Surface Representation for Point Clouds

    Haoxi Ran, Jun Liu, and Chengjie Wang. Surface Representation for Point Clouds. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 18942–18952, June 2022

    2022

  36. [44]

    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies

    Guocheng Qian, Yuchen Li, Houwen Peng, Jinjie Mai, Hasan Hammoud, Mohamed Elhoseiny, and Bernard Ghanem. PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), pages 231...

    2022

  37. [45]

    Augmented Neural ODEs

    Emilien Dupont, Arnaud Doucet, and Yee Whye Teh. Augmented Neural ODEs. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), pages 3140–3150, December 2019

    2019

  38. [46]

    Second-Order Neural ODE Optimizer

    Guan-Horng Liu, Tianrong Chen, and Evangelos Theodorou. Second-Order Neural ODE Optimizer. In Pro- ceedings of the Advances in Neural Information Processing Systems (NeurIPS), pages 25267–25279, December 2021

    2021

  39. [47]

    ANODE: Unconditionally Accurate Memory-Efficient Gradi- ents for Neural ODEs

    Amir Gholami, Kurt Keutzer, and George Biros. ANODE: Unconditionally Accurate Memory-Efficient Gradi- ents for Neural ODEs. In Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI) , pages 730–736, August 2019

    2019

  40. [48]

    Mahoney, and George Biros

    Tianjun Zhang, Zhewei Yao, Amir Gholami, Joseph E Gonzalez, Kurt Keutzer, Michael W. Mahoney, and George Biros. ANODEV2: A Coupled Neural ODE Framework. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), pages 5151–5161, December 2019

    2019

  41. [49]

    Neural SDE: Stabilizing Neural ODE Networks with Stochastic Noise

    Xuanqing Liu, Tesi Xiao, Si Si, Qin Cao, Sanjiv Kumar, and Cho-Jui Hsieh. Neural SDE: Stabilizing Neural ODE Networks with Stochastic Noise. arXiv Preprint 1906.02355, June 2019

    1906 arXiv

  42. [50]

    Graph Neural Ordinary Differential Equations

    Michael Poli, Stefano Massaroli, Junyoung Park, Atsushi Yamashita, Hajime Asama, and Jinkyoo Park. Graph Neural Ordinary Differential Equations. arXiv Preprint 1911.07532, November 2019

    1911 arXiv

  43. [51]

    Continuous Graph Neural Networks

    Louis-Pascal Xhonneux, Meng Qu, and Jian Tang. Continuous Graph Neural Networks. In Proceedings of the International Conference on Machine Learning (ICML), pages 10432–10441, July 2020

    2020

  44. [52]

    Exchangeable Neural ODE for Set Modeling

    Yang Li, Haidong Yi, Christopher Bender, Siyuan Shan, and Junier B Oliva. Exchangeable Neural ODE for Set Modeling. In Proceedings of the Advances in Neural Information Processing Systems (NeurIPS) , pages 6936–6946, December 2020

    2020

  45. [53]

    Deep Sets

    Manzil Zaheer, Satwik Kottur, Siamak Ravanbakhsh, Barnabas Poczos, Ruslan Salakhutdinov, and Alexander J Smola. Deep Sets. In Proceedings of the Advances in Neural Information Processing Systems (NIPS) , pages 3391–3401, December 2017

    2017

  46. [54]

    Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks

    Juho Lee, Yoonho Lee, Jungtaek Kim, Adam Kosiorek, Seungjin Choi, and Yee Whye Teh. Set Transformer: A Framework for Attention-based Permutation-Invariant Neural Networks. In Proceedings of the International Conference on Machine Learning (ICML), pages 3744–3753, July 2019

    2019

  47. [55]

    CaSPR: Learn- ing Canonical Spatiotemporal Point Cloud Representations

    Davis Rempe, Tolga Birdal, Yongheng Zhao, Zan Gojcic, Srinath Sridhar, and Leonidas Guibas. CaSPR: Learn- ing Canonical Spatiotemporal Point Cloud Representations. In Proceedings of the Advances in Neural Informa- tion Processing Systems (NeurIPS), pages 13688–13701, December 2020

    2020

  48. [56]

    A Low-Cost Neural ODE with Depthwise Separable Convolution for Edge Domain Adaptation on FPGAs

    Hiroki Kawakami, Hirohisa Watanabe, Keisuke Sugiura, and Hiroki Matsutani. A Low-Cost Neural ODE with Depthwise Separable Convolution for Edge Domain Adaptation on FPGAs. IEICE Transactions on Information and Systems, E106-D(7):1186–1197, July 2023

    2023

  49. [57]

    A Cost-Efficient FPGA-Based CNN-Transformer Using Neural ODE

    Ikumi Okubo, Keisuke Sugiura, and Hiroki Matsutani. A Cost-Efficient FPGA-Based CNN-Transformer Using Neural ODE. IEEE Access, 12(1):155773–155788, October 2024

    2024

  50. [58]

    Accelerating Neural-ODE Inference on FPGAs with Two-Stage Structured Pruning and History-based Stepsize Search

    Lei Cai, Jing Wang, Lianfeng Yu, Bonan Yan, Yaoyu Tao, and Yuchao Yang. Accelerating Neural-ODE Inference on FPGAs with Two-Stage Structured Pruning and History-based Stepsize Search. In Proceedings of the 2023 ACM/SIGDA International Symposium on Field Programmable Gate Array...

    2023

  51. [59]

    Hardware Implementation of nmODE on FPGA

    Yi Chen, Hanwen Liu, Enqi Zhang, Hong Qu, and Zhang Yi. Hardware Implementation of nmODE on FPGA. In Proceedings of the International Annual Conference on Complex Systems and Intelligent Science (CSIS-IAC), pages 240–246, October 2023

    2023

  52. [60]

    An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An Image is Worth 16x16 Words: Transformers for Image Recognition at...

    2021

  53. [61]

    Qi, Jean-Emmanuel Deschaud, Beatriz Marcotegui, Francois Goulette, and Leonidas J

    Hugues Thomas, Charles R. Qi, Jean-Emmanuel Deschaud, Beatriz Marcotegui, Francois Goulette, and Leonidas J. Guibas. KPConv: Flexible and Deformable Convolution for Point Clouds. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 6411–6420...

    2019

  54. [62]

    Jimmy Lei Ba, Jamie Ryan Kiros, and Geoffrey E. Hinton. Layer Normalization. arXiv Preprint 1607.06450, July 2016

    2016 arXiv

  55. [63]

    3D ShapeNets: A Deep Representation for V olumetric Shapes

    Zhirong Wu, Shuran Song, Aditya Khosla, Fisher Yu, Linguang Zhang, Xiaoou Tang, and Jianxiong Xiao. 3D ShapeNets: A Deep Representation for V olumetric Shapes. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1912–1920, June 2015

    1912

  56. [64]

    Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Data

    Mikaela Angelina Uy, Quang-Hieu Pham, Binh-Son Hua, Thanh Nguyen, and Sai-Kit Yeung. Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Data. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), p...

    2019

Pith tools

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