Pith. sign in

REVIEW 5 major objections 6 minor 97 references

Optimized CNNs for Rapid 3D Point Cloud Object Recognition

T0 review · 5 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read The paper claims that three-layer sparse voting CNNs with L1 regularization beat the prior state of the art on the MVTec 3D-AD point-cloud benchmark.

desk verdict The advertised sparse-voting L1 CNN never appears in the experiments; the paper evaluates a GCN on FPFH + multi-view ResNet features for anomaly detection, making the central claim unsupported. read the letter →

arxiv 2412.02855 v1 pith:E44LPBDY submitted 2024-12-03 cs.CV cs.LG

Tianyi Lyu , Dian Gu , Peiyuan Chen , Yaoting Jiang , Zhenhong Zhang , Huadong Pang , Li Zhou , Yiping Dong This is my paper · ORCID
classification cs.CVcs.LG
keywords 3DpointcloudobjectdetectionsparseconvolutionfeaturevotingL1regularizationMVTec3D-ADanomalygraph
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 claims that convolutional layers built from a feature-centric voting mechanism, combined with an L1 penalty on filter activations, enable a three-layer network to detect objects in sparse 3D point clouds faster and more accurately than prior methods. On the MVTec 3D-AD benchmark, the proposed Vote3Deep models are said to outperform state-of-the-art laser-only and laser-vision fusion approaches while keeping competitive speeds. The significance, if true, is that real-time 3D perception for autonomous driving and robotics could be done natively on point clouds, without projecting to 2D or using dense grids. The full text, however, describes a graph convolutional network that concatenates FPFH and multi-view ResNet features for anomaly detection, so the abstract's mechanism and the reported experiments should be read as separate claims.

What carries the argument

The advertised mechanism is the feature-centric voting rule from Vote3D, which the paper claims can be implemented as sparse convolutional filters that touch only occupied voxels, together with an L1 penalty on activations to enforce intermediate sparsity. In the full text, the operational layer is presented as a graph convolution with $H^{(l)} = \sigma(\tilde{D}^{-1/2}\tilde{A}\tilde{D}^{-1/2}H^{(l-1)}W^{(l)})$, where the adjacency comes from k-nearest neighbors, and no voting or L1 term appears. The graph convolution is what carries the reported experiments; the voting-plus-L1 description carries the abstract's contribution claim.

What would settle it

Train the network exactly as the abstract specifies (a three-layer vote-based sparse CNN with an L1 penalty on activations) on the MVTec 3D-AD data and measure I-ROC and P-PRO; if the resulting scores do not reach 95.15% I-ROC and 92.93% P-PRO, or if the only architecture that reaches those numbers is the FPFH-plus-ResNet graph-convolution network in the full text, then the central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that feature-centric voting can be generalized from the handcrafted Vote3D detector to learned convolutional layers, so that computation scales with occupied cells rather than the full 3D grid, and that adding an L1 penalty on filter activations preserves this sparsity across intermediate layers. A three-layer network built this way, referred to as Vote3Deep, is asserted to achieve state-of-the-art accuracy on the MVTec 3D-AD benchmark, improving average precision by up to 40% over prior laser-based methods while maintaining competitive processing speed. The authors intend this as evidence that native 3D sparse convolution can replace 2D projection and dense volumetric processing in real-time object detection.

Load-bearing premise

The central claim rests on the assumption that the graph convolution in equation (9) is the sparse, feature-voting convolutional layer described in the abstract, and that the L1 penalty is part of the training objective; the paper's equations and experiment section provide no voting rule and no L1 loss term.

Editorial extensions

If this is right

  • If the voting-based sparse layers and L1 regularization work as described, native 3D point-cloud detection could run in real time, with computation proportional only to occupied voxels.
  • A three-layer network replacing much deeper multi-modal pipelines would simplify training, reduce memory footprint, and cut inference latency.
  • L1-induced sparse intermediate activations would lower the cost of later detector stages and make pruning and compression of the model easier.
  • The claimed up-to-40% average-precision gain over prior laser-based methods would set a new benchmark on MVTec 3D-AD and make sparse 3D CNNs attractive for robotics and autonomous driving.

Reading between the lines

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

  • The full text's equations and experiments never exhibit a voting rule or an L1 loss, so the high scores on MVTec 3D-AD may come from the FPFH-plus-ResNet feature concatenation rather than from the promised sparse convolutional layers; a minimal check is to replace the graph convolution with a linear classifier on the same features.
  • If the L1 penalty is absent from the training objective, the paper's proposed sparsity-accuracy trade-off remains unverified; comparing training with and without the penalty on an identical architecture would directly test this.
  • The benchmark is the unsupervised anomaly-detection split of MVTec 3D-AD, whereas the Vote3D baseline targets supervised car/pedestrian/cyclist detection in KITTI-style data; the claimed advance over 'laser-vision' methods may not transfer to supervised 3D object detection suites.
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

5 major / 6 minor

Summary. The paper advertises a method for 3D point cloud object detection based on feature-centric voting sparse convolutional layers and an L1 regularization penalty on filter activations, claiming that three-layer Vote3Deep models outperform prior state of the art on the MVTec 3D-AD benchmark while remaining fast. The actual manuscript describes a pipeline that removes background planes, extracts FPFH and multi-view ResNet18 features, applies a graph convolution (Eq. (9)), pools to a graph-level representation, and scores points with an MLP; Section 4 evaluates I-ROC and P-PRO on the MVTec 3D-AD anomaly-detection benchmark. The central claim and the evaluated method diverge substantially.

Significance. If the claimed three-layer sparse voting CNN with L1-induced sparsity really outperformed prior laser-only and laser-vision methods on a 3D object detection benchmark, that would be a significant practical result for real-time perception. However, the paper as written does not establish this: the only implemented layer is a standard graph convolution, no L1 term appears in the training objective, and the experiments are on unsupervised anomaly detection rather than object detection. The reported anomaly-detection numbers (95.15% I-ROC, 92.93% P-PRO) are competitive, but without variance estimates or a clean table they cannot be adjudicated, and the L1 and voting contributions are untested. On the positive side, the paper uses an external benchmark and explicitly acknowledges limitations in Section 4.3.4, so there is no circular derivation; the problem is that the experiments do not test the advertised method.

major comments (5)
  1. [Section 3.2, Eq. (9)] Equation (9) defines H^(l) = sigma(D^(-1/2) A D^(-1/2) H^(l-1) W^(l)), which is the standard graph-convolution update on a k-NN adjacency matrix. It contains no voting rule, no sparse convolution operator, and no mechanism that restricts computation to occupied cells; the Introduction's statement that the voting mechanism is equivalent to dense convolution and scales with occupied cells is never realized in the architecture. Consequently, the reported results in Section 4 do not evaluate the sparse Vote3Deep layer that the abstract claims.
  2. [Sections 1 and 3, training objective] The abstract and Introduction promise an L1 penalty on filter activations to promote sparsity in intermediate layers, and contribution 2 in Section 1 repeats this claim. However, no training objective, loss function, regularization coefficient, or ablation involving L1 appears anywhere in Sections 3-4. The L1 mechanism is therefore not part of the evaluated system, and the efficiency argument built on it is unsupported.
  3. [Section 4, Tables 1-2] The evaluation is on the MVTec 3D-AD anomaly-detection benchmark with I-ROC and P-PRO metrics, and the model outputs anomaly scores (Eq. (12)) that are thresholded in Eq. (15). This is unsupervised anomaly detection and localization, not 3D object detection with bounding boxes and average precision. The abstract and Introduction frame the contribution as object detection and compare against laser-only and laser-vision detection methods; those claims are not supported by these experiments.
  4. [Table 3] Table 3 is unreadable: rows labeled 'P-PRO 0.8304 0.9230' and 'ResNet18 873±234 0.8918 0.9145' mix what appear to be different feature configurations and numerical values with no row or column labels, no units for 873±234, and no explanation of the two-line entries. In addition, Tables 1 and 2 report no error bars or number of runs, so the claimed improvements over CPMF (I-ROC 0.9515 vs 0.9502; P-PRO 0.9293 vs 0.9282) are within plausible noise and are not statistically established.
  5. [Section 4, implementation details] No runtime, latency, or throughput measurements are reported, despite the abstract's claims of competitive processing speeds and real-time suitability and Section 2's extensive discussion of computational cost. Without timing measurements on the actual model or a complexity analysis of the proposed operations, the efficiency claim is not tested.
minor comments (6)
  1. [Abstract and Section 4.1.1] The benchmark is MVTec 3D-AD anomaly detection, not object detection; the terminology should be corrected throughout, including the abstract, Section 1's 'object detection benchmark', and Section 4.1.1's 'MVTec 3D dataset'.
  2. [Eq. (4)] The equation contains the typo 'ender(P, v)'; it should read 'Render(P, v)'.
  3. [Eq. (14)] The symbol 'Ff uesd(pi)' is a typo for 'F_fused(pi)', and the notation for the fused feature vector is inconsistent between Eqs. (13) and (14) and the surrounding text.
  4. [Table captions] Table 1's caption uses 'I-AUC' while the text and Table 2 use 'I-ROC' and 'P-PRO'; the metric names should be unified.
  5. [Section 4.3.1 and Figures 3-4] The text says 'Fig. 3 displays two examples of images taken from different views', but Figure 3 is the quantitative comparison and the view examples appear in Figure 4; the cross-references are swapped.
  6. [References] Reference [19] is cited as the MVTec 3D-AD Vision Benchmark Suite, but [19] is the KITTI benchmark paper; the MVTec 3D-AD dataset is reference [2].

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation; the advertised Vote3Deep/L1 sparse-convolution method is absent from the implemented pipeline, but the benchmark results are external and not constructed from the target claim.

full rationale

The paper does not contain a circular derivation in the sense of a fitted parameter renamed as a prediction or an equation reducing to its own input by construction. The implemented pipeline is explicit: FPFH features and multi-view ResNet features are concatenated, a k-NN graph is built, graph convolution is applied through Eq. (9) H^(l) = sigma(D^(-1/2) A D^(-1/2) H^(l-1) W^(l)), a readout and MLP produce anomaly scores, and these are evaluated on the external MVTec 3D-AD benchmark with I-ROC and P-PRO metrics. None of these quantities is fitted to the reported I-ROC or P-PRO values, and no uniqueness theorem or self-citation chain forces the claimed outcome. The most serious problem is a method/claim mismatch, not circularity: the abstract promises a feature-centric voting mechanism and L1-regularized sparse convolutional layers, but no voting rule, no sparse-convolution operator, and no L1 loss term appear anywhere in the method. Eq. (9) is the standard graph-convolution rule, and the experiments evaluate anomaly detection rather than 3D object detection. This makes the headline state-of-the-art claim unsupported, but the unsupported claim is not equivalent to its inputs; it is simply untested by the described experiments. There are many self-citations in the reference list, including refs. [8], [14], [39], [42], [43], [52], [61], [83], [86], [91], and [94], but none of them is load-bearing for the central claim: the key external supports are the MVTec dataset, BTF, CPMF, and Vote3D. The limitation subsection honestly acknowledges that some anomalies are only identifiable with RGB information and that rendered-image quality can be degraded by acquisition noise. Overall, no circular step can be exhibited from the paper's equations or citations; the appropriate finding is no significant circularity, with a low score reflecting minor self-citation noise and the substantial gap between the advertised and actual method, which belongs to correctness risk rather than circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper's central claims rest on unstated equivalences and untested components. The principal one is the identification of the graph convolution in Eq. (9) with the abstract's sparse voting convolution; this equivalence is never shown. The L1 penalty is promised but never appears in a loss function, experiment, or hyperparameter. The benchmark itself is anomaly detection, not object detection, so the abstract's framing is unsupported. No code or parameters are provided.

free parameters (4)
  • L1 regularization coefficient = not reported
    The central contribution of the paper is sparsity via L1 penalty, but the coefficient is never stated, so the experiments cannot be reproduced and the contribution is not tested.
  • Number of rendering views Nv = not reported for final results; experiments explore 1,3,6,...,27
    The final I-ROC/P-PRO numbers in Tables 1 and 2 depend on an undisclosed Nv; Table 3 appears to pair varying Nv with backbones but the mapping is garbled.
  • Anomaly threshold tau = not reported
    The decision rule in Eq. (16) requires tau, but no value or selection procedure is given.
  • k in k-NN graph = not reported
    The graph convolution relies on k-nearest neighbors, but k is not specified.
assumptions (5)
  • ad hoc to paper The graph convolution in Eq. (9) is the sparse feature-centric voting convolution described in the abstract.
    Section 3.2 labels the block Sparse Convolutional Neural Networks but Eq. (9) is a standard GCN update with no voting or sparsity mechanism; the paper never proves the equivalence.
  • ad hoc to paper An L1 penalty on filter activations is part of the training objective.
    The introduction and contributions promise L1 regularization, but no loss function, coefficient, or ablation is reported; the experiments never mention it.
  • domain assumption MVTec 3D-AD anomaly detection results are evidence of object detection performance.
    The abstract frames the work as object detection with laser-only and laser-vision methods, while the experiments measure image-level and pixel-level anomaly detection (I-ROC, P-PRO) on MVTec 3D-AD.
  • domain assumption FPFH handcrafted features, as used in BTF, provide sufficient 3D features for the pipeline.
    The method adopts FPFH without learning it and without independent validation in this setting, relying on prior work [25, 56].
  • domain assumption ImageNet-pretrained ResNet features transfer to rendered PCD views.
    The 2D features are assumed to generalize from ImageNet to multi-view renderings of point clouds; no fine-tuning or domain adaptation is reported.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimized CNNs for Rapid 3D Point Cloud Object Recognition." pith.science (2026). https://pith.science/paper/E44LPBDY

@misc{pith2026241202855,
  author       = {Pith},
  title        = {Pith review of: Optimized CNNs for Rapid 3D Point Cloud Object Recognition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/E44LPBDY}},
  note         = {Machine review of arXiv:2412.02855}
}
abstract

This study introduces a method for efficiently detecting objects within 3D point clouds using convolutional neural networks (CNNs). Our approach adopts a unique feature-centric voting mechanism to construct convolutional layers that capitalize on the typical sparsity observed in input data. We explore the trade-off between accuracy and speed across diverse network architectures and advocate for integrating an $\mathcal{L}_1$ penalty on filter activations to augment sparsity within intermediate layers. This research pioneers the proposal of sparse convolutional layers combined with $\mathcal{L}_1$ regularization to effectively handle large-scale 3D data processing. Our method's efficacy is demonstrated on the MVTec 3D-AD object detection benchmark. The Vote3Deep models, with just three layers, outperform the previous state-of-the-art in both laser-only approaches and combined laser-vision methods. Additionally, they maintain competitive processing speeds. This underscores our approach's capability to substantially enhance detection performance while ensuring computational efficiency suitable for real-time applications.

Figures

Figures reproduced from arXiv: 2412.02855 by the authors.

Figure 1
Figure 1. Our architecture. from various factors such as sensor inaccuracies or en￾vironmental interference. We filter out points with NaN (Not a Number) values or those that do not be￾long to any significant structure. This step ensures that only meaningful data points are retained for further processing: Pclean = P/Abg (2) here, Pclean represents the cleaned point cloud after removing background points Abg. Fast Point Featu… view at source ↗
Figure 2
Figure 2. Visualization of prediction results [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Comparisons of the anomaly detection performances under different types of features, views, and backbones. 7 [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Samples for the influence of different views. views is between approximately 12 and 18. The overall improvements can be clearly attributed to the fact that images from a greater number of views provide a more comprehensive description and capture of the information und…
Figure 6
Figure 6. Figure 6: Illustration for the bad quality of rendered images resulting by acquisition noise. poorly in categories like bagel and potato, where F3D excels. The combination of F2D and F3D, forming FOurs, markedly enhances anomaly detection perfor￾mance and achieves impressive loc…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

97 extracted references · 55 canonical work pages

  1. [1]

    Johnson, Jonathan Sprinkle, and Meiyi Ma

    Ziyan An, Xia Wang, Taylor T. Johnson, Jonathan Sprinkle, and Meiyi Ma. Runtime monitoring of acci- dents in driving recordings with multi-type logic in 10 IECE T ransactions on Internet of Things empirical models. InInternational Conference on Run- time Verification, pages 376–388. Springer, 2023

    2023

  2. [2]

    The mvtec 3d-ad dataset for unsupervised 3d anomaly detection and localization.arXiv preprint arXiv:2112.09045, 2021

    PaulBergmann,XinJin,DavidSattlegger,andCarsten Steger. The mvtec 3d-ad dataset for unsupervised 3d anomaly detection and localization.arXiv preprint arXiv:2112.09045, 2021

    arXiv 2021

  3. [3]

    Anomaly detec- tion in 3d point clouds using deep geometric descrip- tors

    Paul Bergmann and David Sattlegger. Anomaly detec- tion in 3d point clouds using deep geometric descrip- tors. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 2613–2623, 2023

    2023

  4. [4]

    Expectedfile-deliverytimeofdtnprotocolover asymmetric space internetwork channels

    Bin Cao, Ruhai Wang, Alaa Sabbagh, Siwei Peng, Kan- glian Zhao, Juan A Fraire, Guannan Yang, and Yue Wang. Expectedfile-deliverytimeofdtnprotocolover asymmetric space internetwork channels. In2018 6th IEEE International Conference on Wireless for Space and Extreme Environments (WiSEE), pages 147–151. IEEE, 2018

    2018

  5. [5]

    Complementary pseudo multimodal feature for point cloud anomaly detection.Pattern Recognition, 156:110761, 2024

    Yunkang Cao, Xiaohao Xu, and Weiming Shen. Complementary pseudo multimodal feature for point cloud anomaly detection.Pattern Recognition, 156:110761, 2024

    2024

  6. [6]

    Convolutional neural network (cnn) for im- agedetectionandrecognition

    Rahul Chauhan, Kamal Kumar Ghanshala, and RC Joshi. Convolutional neural network (cnn) for im- agedetectionandrecognition. In 2018firstinternational conference on secure cyber computing and communication (ICSCCC), pages 278–282. IEEE, 2018

    2018

  7. [7]

    VoxResNet: Deep Voxelwise Residual Networks for Volumetric Brain Segmentation

    Hao Chen, Qi Dou, Lequan Yu, and Pheng-Ann Heng. Voxresnet: Deep voxelwise residual networks for volumetric brain segmentation. arXiv preprint arXiv:1608.05895, 2016

    work page Pith review arXiv 2016

  8. [8]

    Enhancing visual question answer- ing through ranking-based hybrid training and mul- timodal fusion

    PeiyuanChen,ZechengZhang,YipingDong,LiZhou, and Han Wang. Enhancing visual question answer- ing through ranking-based hybrid training and mul- timodal fusion. Journal of Intelligence Technology and Innovation, 2(3):19–46, 2024

    2024

Show all 97 references
  1. [9]

    Few-shot name entity recognition on stackoverflow

    XinweiChen,KunLi,TianyouSong,andJiangjianGuo. Few-shot name entity recognition on stackoverflow. arXiv preprint arXiv:2404.09405, 2024

    2024 arXiv

  2. [10]

    Mix of experts language model for named entity recognition

    Xinwei Chen, Kun Li, Tianyou Song, and Jiangjian Guo. Mix of experts language model for named entity recognition. arXiv preprint arXiv:2404.19192, 2024

    2024 arXiv

  3. [11]

    Modeling and simulation of dna origami based elec- tronic read-only memory

    Arpan De, Hashem Mohammad, Yiren Wang, Rajku- mar Kubendran, Arindam K Das, and MP Anantram. Modeling and simulation of dna origami based elec- tronic read-only memory. In2022 IEEE 22nd Inter- national Conference on Nanotechnology (NANO), pages 385–388. IEEE, 2022

    2022

  4. [12]

    Performance analysis of dna crossbar arrays for high- densitymemorystorageapplications

    Arpan De, Hashem Mohammad, Yiren Wang, Rajku- mar Kubendran, Arindam K Das, and MP Anantram. Performance analysis of dna crossbar arrays for high- densitymemorystorageapplications. ScientificReports, 13(1):6650, 2023

    2023

  5. [13]

    Overview of the ransac algorithm

    Konstantinos G Derpanis. Overview of the ransac algorithm. Image Rochester NY, 4(1):2–3, 2010

    2010

  6. [14]

    The design of autonomous uav proto- types for inspecting tunnel construction environment

    Yiping Dong. The design of autonomous uav proto- types for inspecting tunnel construction environment. Journal of Intelligence Technology and Innovation, 2(3):1– 18, 2024

    2024

  7. [15]

    Automatic detection of cerebral microbleeds from mr images via 3d convolutional neural networks

    QiDou,HaoChen,LequanYu,LeiZhao,JingQin,De- feng Wang, Vincent CT Mok, Lin Shi, and Pheng-Ann Heng. Automatic detection of cerebral microbleeds from mr images via 3d convolutional neural networks. IEEE transactions on medical imaging, 35(5):1182–1195, 2016

    2016

  8. [16]

    A density-based algorithm for discovering clusters in large spatial databases with noise

    MartinEster,Hans-PeterKriegel,JörgSander,Xiaowei Xu, et al. A density-based algorithm for discovering clusters in large spatial databases with noise. Inkdd, volume 96, pages 226–231, 1996

    1996

  9. [17]

    An image detection technique based on morphologi- cal edge detection and background differencing for real-time traffic analysis.Pattern Recognition Letters, 16(12):1321–1330, 1995

    Mahmood Fathy and Mohammed Yakoob Siyal. An image detection technique based on morphologi- cal edge detection and background differencing for real-time traffic analysis.Pattern Recognition Letters, 16(12):1321–1330, 1995

    1995

  10. [18]

    Random sam- pleconsensus: aparadigmformodelfittingwithappli- cations to image analysis and automated cartography

    Martin A Fischler and Robert C Bolles. Random sam- pleconsensus: aparadigmformodelfittingwithappli- cations to image analysis and automated cartography. Communications of the ACM, 24(6):381–395, 1981

    1981

  11. [19]

    Are we ready for autonomous driving? the kitti vision benchmark suite

    AndreasGeiger,PhilipLenz,andRaquelUrtasun. Are we ready for autonomous driving? the kitti vision benchmark suite. In2012 IEEE conference on computer vision and pattern recognition, pages 3354–3361. IEEE, 2012

    2012

  12. [20]

    Research on empirical correction models of gps block iif and bds satellite inter-frequency clock bias

    Xiaopeng Gong, Shengfeng Gu, Yidong Lou, Fu Zheng, Xinhao Yang, Zhipeng Wang, and Jingnan Liu. Research on empirical correction models of gps block iif and bds satellite inter-frequency clock bias. Journal of Geodesy, 94:1–11, 2020

    2020

  13. [21]

    Graphical Structural Learning of rs-fMRI data in Heavy Smokers.arXiv preprint arXiv:2409.08395, 2024

    Yiru Gong, Qimin Zhang, Huili Zheng, Zheyan Liu, and Shaohan Chen. Graphical Structural Learning of rs-fMRI data in Heavy Smokers.arXiv preprint arXiv:2409.08395, 2024

    2024 arXiv

  14. [22]

    Sparse3dconvolutionalneuralnetworks

    BenGraham. Sparse3dconvolutionalneuralnetworks. arXiv preprint arXiv:1505.02890, 2015

    2015 arXiv

  15. [23]

    Spatially-sparseconvolutionalneu- ral networks.arXiv preprint arXiv:1409.6070, 2014

    BenjaminGraham. Spatially-sparseconvolutionalneu- ral networks.arXiv preprint arXiv:1409.6070, 2014

    2014 arXiv

  16. [24]

    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, pages 770–778, 2016

    2016

  17. [25]

    Backtothefeature: classical 3d features are (almost) all you need for 3d anomaly detection

    EliahuHorwitzandYedidHoshen. Backtothefeature: classical 3d features are (almost) all you need for 3d anomaly detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 2967–2976, 2023

    2023

  18. [26]

    Relation networks for object detection

    Han Hu, Jiayuan Gu, Zheng Zhang, Jifeng Dai, and Yichen Wei. Relation networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3588–3597, 2018

    2018

  19. [27]

    11 IECE T ransactions on Internet of Things Risk analysis in customer relationship management viaqrcnn-lstmandcross-attentionmechanism

    Yaowen Huang, Jun Der Leu, Baoli Lu, and Yan Zhou. 11 IECE T ransactions on Internet of Things Risk analysis in customer relationship management viaqrcnn-lstmandcross-attentionmechanism. Journal of Organizational and End User Computing (JOEUC), 36(1):1–22, 2024

    2024

  20. [28]

    Learning sparse high dimensional filters: Image fil- tering, dense crfs and bilateral neural networks

    Varun Jampani, Martin Kiefel, and Peter V Gehler. Learning sparse high dimensional filters: Image fil- tering, dense crfs and bilateral neural networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4452–4461, 2016

    2016

  21. [29]

    Trajectory track- ing using frenet coordinates with deep deterministic policy gradient.arXiv preprint arXiv:2411.13885, 2024

    Tongzhou Jiang, Lipeng Liu, Junyue Jiang, Tianyao Zheng, Yuhui Jin, and Kunpeng Xu. Trajectory track- ing using frenet coordinates with deep deterministic policy gradient.arXiv preprint arXiv:2411.13885, 2024

    2024 arXiv

  22. [30]

    Du- alvd: An adaptive dual encoding model for deep vi- sual understanding in visual dialogue

    Xiaoze Jiang, Jing Yu, Zengchang Qin, Yingying Zhuang, Xingxing Zhang, Yue Hu, and Qi Wu. Du- alvd: An adaptive dual encoding model for deep vi- sual understanding in visual dialogue. InProceedings oftheAAAIconferenceonartificialintelligence ,volume34, pages 11125–11132, 2020

    2020

  23. [31]

    Dbscan: Past, present and future

    Kamran Khan, Saif Ur Rehman, Kamran Aziz, Simon Fong, and Sababady Sarasvady. Dbscan: Past, present and future. InThe fifth international conference on the applications of digital information and web technologies (ICADIWT 2014), pages 232–238. IEEE, 2014

    2014

  24. [32]

    Imagenet classification with deep convolutional neuralnetworks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hin- ton. Imagenet classification with deep convolutional neuralnetworks. Advancesinneuralinformationprocess- ing systems, 25, 2012

    2012

  25. [33]

    Traffic smoothing via connected & automated vehicles: A modular, hierarchical control design deployed in a 100-cav flow smoothing experiment.IEEE Control Sys- tems Magazine, 2024

    JWLee,HWang,KJang,AHayat,MBunting,AAlan- qary, WBarbour, ZFu, XGong, GGunter, etal. Traffic smoothing via connected & automated vehicles: A modular, hierarchical control design deployed in a 100-cav flow smoothing experiment.IEEE Control Sys- tems Magazine, 2024

    2024

  26. [34]

    Vehicle detection from3dlidarusingfullyconvolutionalnetwork

    Bo Li, Tianlei Zhang, and Tian Xia. Vehicle detection from3dlidarusingfullyconvolutionalnetwork. arXiv preprint arXiv:1608.07916, 2016

    2016 arXiv

  27. [35]

    Deep reinforcement learning-based obstacle avoidance for robot move- ment in warehouse environments

    Keqin Li, Jiajing Chen, Denzhi Yu, Tao Dajun, Xinyu Qiu, Lian Jieting, Sun Baiwei, Zhang Shengyuan, Zhenyu Wan, Ran Ji, et al. Deep reinforcement learning-based obstacle avoidance for robot move- ment in warehouse environments. arXiv preprint arXiv:2409.14972, 2024

    2024 arXiv

  28. [36]

    Optimizing automated picking systems in warehouse robots using machine learning

    Keqin Li, Jin Wang, Xubo Wu, Xirui Peng, Runmian Chang, Xiaoyu Deng, Yiwen Kang, Yue Yang, Fang- hao Ni, and Bo Hong. Optimizing automated picking systems in warehouse robots using machine learning. arXiv preprint arXiv:2408.16633, 2024

    2024

  29. [37]

    Ltpnet inte- gration of deep learning and environmental decision support systems for renewable energy demand fore- casting

    Te Li, Mengze Zhang, and Yan Zhou. Ltpnet inte- gration of deep learning and environmental decision support systems for renewable energy demand fore- casting. arXiv preprint arXiv:2410.15286, 2024

    2024 arXiv

  30. [38]

    Enhancing the reliability of out-of-distribution im- age detection in neural networks

    Shiyu Liang, Yixuan Li, and Rayadurgam Srikant. Enhancing the reliability of out-of-distribution im- age detection in neural networks. arXiv preprint arXiv:1706.02690, 2017

    2017 arXiv

  31. [39]

    Dsem- nerf: Multimodalfeaturefusionandglobal-localatten- tion for enhanced 3d scene reconstruction.Information Fusion, page 102752, 2024

    Dong Liu, Zhiyong Wang, and Peiyuan Chen. Dsem- nerf: Multimodalfeaturefusionandglobal-localatten- tion for enhanced 3d scene reconstruction.Information Fusion, page 102752, 2024

    2024

  32. [40]

    Guiran Liu and Binrong Zhu. Design and implemen- tation of intelligent robot control system integrating computer vision and mechanical engineering.Interna- tional Journal of Computer Science and Information Tech- nology, 3(1):219–226, 2024

    2024

  33. [41]

    Td3 based collision free mo- tion planning for robot navigation

    Hao Liu, Yi Shen, Chang Zhou, Yuelin Zou, Zijun Gao, and Qi Wang. Td3 based collision free mo- tion planning for robot navigation. arXiv preprint arXiv:2405.15460, 2024

    2024 arXiv

  34. [42]

    Eitnet: An iot-enhanced framework for real-time bas- ketball action recognition.Alexandria Engineering Jour- nal, 110:567–578, 2025

    Jingyu Liu, Xinyu Liu, Mingzhe Qu, and Tianyi Lyu. Eitnet: An iot-enhanced framework for real-time bas- ketball action recognition.Alexandria Engineering Jour- nal, 110:567–578, 2025

    2025

  35. [43]

    Real-time monitor- ing of lower limb movement resistance based on deep learning

    Yuanmeng Liu, Tianyi Lyu, et al. Real-time monitor- ing of lower limb movement resistance based on deep learning. Alexandria Engineering Journal, 111:136–147, 2025

    2025

  36. [44]

    Fleet rebalancing for expanding shared e- mobility systems: A multi-agent deep reinforcement learning approach

    Man Luo, Bowen Du, Wenzhe Zhang, Tianyou Song, Kun Li, Hongming Zhu, Mark Birkin, and Hongkai Wen. Fleet rebalancing for expanding shared e- mobility systems: A multi-agent deep reinforcement learning approach. IEEE Transactions on Intelligent Transportation Systems, 24(4):386...

    2023

  37. [45]

    3d convolu- tionalneuralnetworksforlandingzonedetectionfrom lidar

    Daniel Maturana and Sebastian Scherer. 3d convolu- tionalneuralnetworksforlandingzonedetectionfrom lidar. In2015 IEEE international conference on robotics and automation (ICRA), pages 3471–3478. IEEE, 2015

    2015

  38. [46]

    Voxnet: A 3d convolutional neural network for real-time object recognition

    Daniel Maturana and Sebastian Scherer. Voxnet: A 3d convolutional neural network for real-time object recognition. In2015 IEEE/RSJ international conference on intelligent robots and systems (IROS), pages 922–928. IEEE, 2015

    2015

  39. [47]

    3d object detection with pointformer

    Xuran Pan, Zhuofan Xia, Shiji Song, Li Erran Li, and Gao Huang. 3d object detection with pointformer. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7463–7472, 2021

    2021

  40. [48]

    Maxk-gnn: Extremely fast gpu kernel design for accelerating graph neural net- works training

    Hongwu Peng, Xi Xie, Kaustubh Shivdikar, Md Amit Hasan,JiahuiZhao,ShaoyiHuang,OmerKhan,David Kaeli, and Caiwen Ding. Maxk-gnn: Extremely fast gpu kernel design for accelerating graph neural net- works training. InProceedings of the 29th ACM Interna- tional Conference on Archit...

    2024

  41. [49]

    Automatic news gen- eration and fact-checking system based on language processing

    Xirui Peng, Qiming Xu, Zheng Feng, Haopeng Zhao, Lianghao Tan, Yan Zhou, Zecheng Zhang, Chenwei Gong, and Yingqiao Zheng. Automatic news gen- eration and fact-checking system based on language processing. JournalofIndustrialEngineeringandApplied Science, 2(3):1–11, 2024

    2024

  42. [50]

    Deep feature learning for knee cartilage segmentation us- ing a triplanar convolutional neural network

    Adhish Prasoon, Kersten Petersen, Christian Igel, 12 IECE T ransactions on Internet of Things François Lauze, Erik Dam, and Mads Nielsen. Deep feature learning for knee cartilage segmentation us- ing a triplanar convolutional neural network. InIn- ternational conference on med...

    2013

  43. [51]

    Robust domain generalization for multi-modal object recog- nition

    Yuxin Qiao, Keqin Li, Junhong Lin, Rong Wei, Chufeng Jiang, Yang Luo, and Haoyu Yang. Robust domain generalization for multi-modal object recog- nition. In2024 5th International Conference on Artificial Intelligence and Electromechanical Automation (AIEA), pages 392–397. IEEE, 2024

    2024

  44. [52]

    Iot-based 3d pose estimation and motion optimization for athletes: Application of c3d and openpose

    Fei Ren, Chao Ren, and Tianyi Lyu. Iot-based 3d pose estimation and motion optimization for athletes: Application of c3d and openpose. arXiv preprint arXiv:2411.12676, 2024

    2024 arXiv

  45. [53]

    Reinforcement learning with com- munication latency with application to stop-and-go wave dissipation

    Alex Richardson, Xia Wang, Abhishek Dubey, and Jonathan Sprinkle. Reinforcement learning with com- munication latency with application to stop-and-go wave dissipation. In2024 IEEE Intelligent Vehicles Sym- posium (IV), pages 1187–1193. IEEE, 2024

    2024

  46. [54]

    Asymmetricstudent-teachernetworks for industrial anomaly detection

    MarcoRudolph,TomWehrbein,BodoRosenhahn,and BastianWandt. Asymmetricstudent-teachernetworks for industrial anomaly detection. InProceedings of the IEEE/CVF winter conference on applications of computer vision, pages 2592–2602, 2023

    2023

  47. [55]

    Im- agenet large scale visual recognition challenge.Inter- national journal of computer vision, 115:211–252, 2015

    OlgaRussakovsky,JiaDeng,HaoSu,JonathanKrause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Im- agenet large scale visual recognition challenge.Inter- national journal of computer vision, 115:211–252, 2015

    2015

  48. [56]

    Fast point feature histograms (fpfh) for 3d registra- tion

    Radu Bogdan Rusu, Nico Blodow, and Michael Beetz. Fast point feature histograms (fpfh) for 3d registra- tion. In 2009 IEEE international conference on robotics and automation, pages 3212–3217. IEEE, 2009

    2009

  49. [57]

    Harnessing XGBoost for robust biomarker selec- tion of obsessive-compulsive disorder (OCD) from adolescentbraincognitivedevelopment(ABCD)data

    Xinyu Shen, Qimin Zhang, Huili Zheng, and Weiwei Qi. Harnessing XGBoost for robust biomarker selec- tion of obsessive-compulsive disorder (OCD) from adolescentbraincognitivedevelopment(ABCD)data. In Pier Paolo Piccaluga, Ahmed El-Hashash, and Xi- angqian Guo, editors,Fourth In...

    2024

  50. [58]

    Deep learning powered estimate of the extrinsic parameters on unmanned surface ve- hicles

    Yi Shen, Hao Liu, Chang Zhou, Wentao Wang, Zijun Gao, and Qi Wang. Deep learning powered estimate of the extrinsic parameters on unmanned surface ve- hicles. arXiv preprint arXiv:2406.04821, 2024

    2024 arXiv

  51. [59]

    Multi-gnss satellite clock estimation con- strained with oscillator noise model in the existence of data discontinuity.Journal of Geodesy, 93:515–528, 2019

    Chuang Shi, Shiwei Guo, Shengfeng Gu, Xinhao Yang, Xiaopeng Gong, Zhiguo Deng, Maorong Ge, and Har- ald Schuh. Multi-gnss satellite clock estimation con- strained with oscillator noise model in the existence of data discontinuity.Journal of Geodesy, 93:515–528, 2019

    2019

  52. [60]

    Very deep convolutional networks for large-scale image recogni- tion

    Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recogni- tion. arXiv preprint arXiv:1409.1556, 2014

    2014 arXiv

  53. [61]

    Mingxiu Sui, Liheng Jiang, Tianyi Lyu, Han Wang, Li Zhou, Peiyuan Chen, and Ammar Alhosain. Appli- cation of deep learning models based on efficientdet and openpose in user-oriented motion rehabilitation robot control.Journal of Intelligence Technology and In- novation, 2(3):4...

    2024

  54. [62]

    Support vec- tor machine.Machine learning models and algorithms for big data classification: thinking with examples for effective learning, pages 207–235, 2016

    Shan Suthaharan and Shan Suthaharan. Support vec- tor machine.Machine learning models and algorithms for big data classification: thinking with examples for effective learning, pages 207–235, 2016

    2016

  55. [63]

    Real-time mon- itoring and analysis of track and field athletes based on edge computing and deep reinforcement learning algorithm

    Xiaowei Tang, Bin Long, and Li Zhou. Real-time mon- itoring and analysis of track and field athletes based on edge computing and deep reinforcement learning algorithm. arXiv preprint arXiv:2411.06720, 2024

    2024 arXiv

  56. [64]

    Image anomaly detection and prediction scheme based on ssa optimized resnet50- bigru model.arXiv preprint arXiv:2406.13987, 2024

    Qianhui Wan, Zecheng Zhang, Liheng Jiang, Zhaoqi Wang, and Yan Zhou. Image anomaly detection and prediction scheme based on ssa optimized resnet50- bigru model.arXiv preprint arXiv:2406.13987, 2024

    2024 arXiv

  57. [65]

    Theoretical analysis of meta reinforcement learning: Generalization bounds and convergence guarantees

    Cangqing Wang, Mingxiu Sui, Dan Sun, Zecheng Zhang, and Yan Zhou. Theoretical analysis of meta reinforcement learning: Generalization bounds and convergence guarantees. CMNM ’24, page 153–159, New York, NY, USA, 2024. Association for Computing Machinery

    2024

  58. [66]

    Voting for voting in online point cloud object detection

    Dominic Zeng Wang and Ingmar Posner. Voting for voting in online point cloud object detection. In Robotics: science and systems, volume 1, pages 10–15. Rome, Italy, 2015

    2015

  59. [67]

    Jingyi Wang, Zhiqun Wang, and Guiran Liu. Record- ing brain activity while listening to music using wear- able eeg devices combined with bidirectional long short-term memory networks.Alexandria Engineering Journal, 109:1–10, 2024

    2024

  60. [68]

    Cross-border commodity pricing strategy optimization via mixed neural network for time series analysis.arXiv preprint arXiv:2408.12115, 2024

    Lijuan Wang, Yijia Hu, and Yan Zhou. Cross-border commodity pricing strategy optimization via mixed neural network for time series analysis.arXiv preprint arXiv:2408.12115, 2024

    2024 arXiv

  61. [69]

    Intelligent design and optimiza- tion of exercise equipment based on fusion algorithm of yolov5-resnet 50

    Long Wang, Wendong Ji, Gang Wang, Yinqiu Feng, and Minghua Du. Intelligent design and optimiza- tion of exercise equipment based on fusion algorithm of yolov5-resnet 50. Alexandria Engineering Journal, 104:710–722, 2024

    2024

  62. [70]

    Deeplearning-basedanomalydetectionandlog analysis for computer networks.Journal of Information and Computing, 2(2):34–63, 2024

    Shuzhan Wang, Ruxue Jiang, Zhaoqi Wang, and Yan Zhou. Deeplearning-basedanomalydetectionandlog analysis for computer networks.Journal of Information and Computing, 2(2):34–63, 2024

    2024

  63. [71]

    Using automated vehicle data as a fitness tracker for sustainability

    Xia Wang, Sobenna Onwumelu, and Jonathan Sprin- kle. Using automated vehicle data as a fitness tracker for sustainability. In2024 Forum for Innovative Sustain- able Transportation Systems (FISTS), pages 1–6. IEEE, 2024

    2024

  64. [72]

    A machine learn- ing approach for accurate and real-time dna sequence identification

    Yiren Wang, Mashari Alangari, Joshua Hihath, Arindam K Das, and MP Anantram. A machine learn- ing approach for accurate and real-time dna sequence identification. BMC genomics, 22:1–10, 2021. 13 IECE T ransactions on Internet of Things

    2021

  65. [73]

    Computational study of the role of counterions and solvent dielectric in determining the conductance of b-dna.Physical Review E, 107(4):044404, 2023

    Yiren Wang, Busra Demir, Hashem Mohammad, Ersin Emre Oren, and MP Anantram. Computational study of the role of counterions and solvent dielectric in determining the conductance of b-dna.Physical Review E, 107(4):044404, 2023

    2023

  66. [74]

    Classification of dna sequences: Performance evaluation of multiple machine learning methods

    Yiren Wang, Vikram Khandelwal, Arindam K Das, and MP Anantram. Classification of dna sequences: Performance evaluation of multiple machine learning methods. In2022IEEE22ndInternationalConferenceon Nanotechnology (NANO), pages 333–336. IEEE, 2022

    2022

  67. [75]

    Performance evaluation of quic with bbr in satellite internet

    Yue Wang, Kanglian Zhao, Wenfeng Li, Juan Fraire, Zhili Sun, and Yuan Fang. Performance evaluation of quic with bbr in satellite internet. In2018 6th IEEE InternationalConferenceonWirelessforSpaceandExtreme Environments (WiSEE), pages 195–199. IEEE, 2018

    2018

  68. [76]

    View-gcn: View- based graph convolutional network for 3d shape anal- ysis

    Xin Wei, Ruixuan Yu, and Jian Sun. View-gcn: View- based graph convolutional network for 3d shape anal- ysis. InProceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, pages 1850–1859, 2020

    2020

  69. [77]

    Big data and machine learning in defence

    Yijie Weng. Big data and machine learning in defence. InternationalJournalofComputerScienceandInformation Technology, 16(2):25–35, 2024

    2024

  70. [78]

    Leveraging artificial in- telligence to enhance data security and combat cyber attacks

    Yijie Weng and Jianhao Wu. Leveraging artificial in- telligence to enhance data security and combat cyber attacks. Journal of Artificial Intelligence General science (JAIGS) ISSN: 3006-4023, 5(1):392–399, 2024

    2024

  71. [79]

    Yijie Weng, Jianhao Wu, et al. Fortifying the global data fortress: a multidimensional examination of cy- ber security indexes and data protection measures across 193 nations.International Journal of Frontiers in Engineering Technology, 6(2):13–28, 2024

    2024

  72. [80]

    Comprehensive overview of artificial intelligence applications in modern industries

    Yijie Weng, Jianhao Wu, Tara Kelly, and William John- son. Comprehensive overview of artificial intelligence applications in modern industries. arXiv preprint arXiv:2409.13059, 2024

    2024 arXiv

  73. [81]

    Xinyao Xi, Chen Zhang, Wen Jia, and Ruxue Jiang. Enhancing human pose estimation in sports training: Integrating spatiotemporal transformer for improved accuracy and real-time performance.Alexandria Engi- neering Journal, 109:144–156, 2024

    2024

  74. [82]

    Accel-gcn: High- performance gpu accelerator design for graph con- volution networks

    Xi Xie, Hongwu Peng, Amit Hasan, Shaoyi Huang, Jiahui Zhao, Haowen Fang, Wei Zhang, Tong Geng, Omer Khan, and Caiwen Ding. Accel-gcn: High- performance gpu accelerator design for graph con- volution networks. In2023 IEEE/ACM International Conference on Computer Aided Design (I...

    2023

  75. [83]

    Dpmpc-planner: A real-time uav trajectory planning framework for complex static environments with dy- namic obstacles

    ZhefanXu,DiDeng,YipingDong,andKenjiShimada. Dpmpc-planner: A real-time uav trajectory planning framework for complex static environments with dy- namic obstacles. In 2022 International Conference on Robotics and Automation (ICRA), pages 250–256. IEEE, 2022

    2022

  76. [84]

    Application of deep learning for automatic identifica- tion of hazardous materials and urban safety supervi- sion

    Tieyi Yan, Jiaxi Wu, Munish Kumar, and Yan Zhou. Application of deep learning for automatic identifica- tion of hazardous materials and urban safety supervi- sion. Journal of Organizational and End User Computing (JOEUC), 36(1):1–20, 2024

    2024

  77. [85]

    Regional bds satellite clock estimation with triple-frequency ambi- guity resolution based on undifferenced observation

    Xinhao Yang, Shengfeng Gu, Xiaopeng Gong, Weiwei Song, Yidong Lou, and Jingnan Liu. Regional bds satellite clock estimation with triple-frequency ambi- guity resolution based on undifferenced observation. GPS Solutions, 23:1–11, 2019

    2019

  78. [86]

    Gta-net: An iot-integrated 3dhumanposeestimationsystemforreal-timeadoles- cent sports posture correction.Alexandria Engineering Journal, 112:585–597, 2025

    Shizhe Yuan and Li Zhou. Gta-net: An iot-integrated 3dhumanposeestimationsystemforreal-timeadoles- cent sports posture correction.Alexandria Engineering Journal, 112:585–597, 2025

    2025

  79. [87]

    Wide resid- ual networks.arXiv preprint arXiv:1605.07146, 2016

    Sergey Zagoruyko and Nikos Komodakis. Wide resid- ual networks.arXiv preprint arXiv:1605.07146, 2016

    2016 arXiv

  80. [88]

    Cu-net: a u-net architecture for efficient brain- tumor segmentation on brats 2019 dataset

    Qimin Zhang, Weiwei Qi, Huili Zheng, and Xinyu Shen. Cu-net: a u-net architecture for efficient brain- tumor segmentation on brats 2019 dataset. arXiv preprint arXiv:2406.13113, 2024

    2019 arXiv

  81. [89]

    Deep analysis of time series data for smart grid startup strategies: A transformer-lstm-pso model approach

    Zecheng Zhang. Deep analysis of time series data for smart grid startup strategies: A transformer-lstm-pso model approach. Journal of Management Science and Operations, 2(3):16–43, 2024

    2024

  82. [90]

    Identification of prognostic biomarkers for stage iii non-small cell lung carcinoma in female nonsmokers using machine learning.arXiv preprint arXiv:2408.16068, 2024

    Huili Zheng, Qimin Zhang, Yiru Gong, Zheyan Liu, and Shaohan Chen. Identification of prognostic biomarkers for stage iii non-small cell lung carcinoma in female nonsmokers using machine learning.arXiv preprint arXiv:2408.16068, 2024

    2024 arXiv

  83. [91]

    Triz method for urban building energy opti- mization: Gwo-sarima-lstmforecastingmodel

    Shirong Zheng, Shaobo Liu, Zhenhong Zhang, Dian Gu, Chunqiu Xia, Huadong Pang, and Enock Mintah Ampaw. Triz method for urban building energy opti- mization: Gwo-sarima-lstmforecastingmodel. Journal of Intelligence Technology and Innovation, 2(3):78–103, 2024

    2024

  84. [92]

    Open3d: A modern library for 3d data processing

    Qian-Yi Zhou, Jaesik Park, and Vladlen Koltun. Open3d: A modern library for 3d data processing. arXiv preprint arXiv:1801.09847, 2018

    2018 arXiv

  85. [93]

    Adapi: Facilitating dnn modeladaptivityforefficientprivateinferenceinedge computing

    TongZhou,JiahuiZhao,YukuiLuo,XiXie,WujieWen, Caiwen Ding, and Xiaolin Xu. Adapi: Facilitating dnn modeladaptivityforefficientprivateinferenceinedge computing. arXiv preprint arXiv:2407.05633, 2024

    2024 arXiv

  86. [94]

    Optimizationofautomatedgarbagerecognitionmodel basedonresnet-50andweaklysupervisedcnnforsus- tainable urban development.Alexandria Engineering Journal, 108:415–427, 2024

    Yan Zhou, Zhaoqi Wang, Shirong Zheng, Li Zhou, Lu Dai, Hao Luo, Zecheng Zhang, and Mingxiu Sui. Optimizationofautomatedgarbagerecognitionmodel basedonresnet-50andweaklysupervisedcnnforsus- tainable urban development.Alexandria Engineering Journal, 108:415–427, 2024

    2024

  87. [95]

    Fa- cial sentiment classification based on resnet-18 model

    YitaoZhou,FujiRen,ShunNishide,andXinKang. Fa- cial sentiment classification based on resnet-18 model. In2019 International Conference on electronic engineering and informatics (EEI), pages 463–466. IEEE, 2019

    2019

  88. [96]

    Complex scene under- standing and object detection algorithm assisted by artificial intelligence.Academic Journal of Science and Technology, 12(3):12–15, 2024

    Binrong Zhu and Guiran Liu. Complex scene under- standing and object detection algorithm assisted by artificial intelligence.Academic Journal of Science and Technology, 12(3):12–15, 2024

    2024

  89. [97]

    In Proceedings of the 2020 4th international conference on digital signal processing, pages 155–159, 2020

    YingyingZhuang,YuezhangChen,andJieZheng.Mu- 14 IECE T ransactions on Internet of Things sic genre classification with transformer classifier. In Proceedings of the 2020 4th international conference on digital signal processing, pages 155–159, 2020. 15

    2020

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.