Pith. sign in

REVIEW 5 major objections 6 minor 35 references

DAGNet: A Dual-View Attention-Guided Network for Efficient X-ray Security Inspection

T0 review · 5 major / 6 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read DAGNet is a dual-view X-ray contraband classifier whose three modules—frequency-domain interaction, hierarchical cross-attention, and guided fusion—raise test mAP by 1.5–2.4 points over AHCR and up to 4.8 points over a dual-view baseline…

desk verdict A plausible modular architecture for dual-view X-ray screening, undermined by sloppy reporting and single-split evidence; worth reviewing but not accepting as-is. read the letter →

arxiv 2502.01710 v4 pith:B4O2G7DL submitted 2025-02-03 cs.CV

Shilong Hong , Yanzhou Zhou , Weichao Xu This is my paper · ORCID
classification cs.CV
keywords dual-viewX-rayinspectionmulti-labelclassificationfrequencydomaininteractioncross-viewattentionfeaturefusioncontrabanddetectionDvXraydatasetmechanism
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

Dual-view X-ray scanners photograph luggage from two directions, but simply feeding both views to a network leaves complementary information unused. This paper argues that three coordinated modules—an adaptive filter in the frequency domain, cross-attention alignment between views, and convolution-guided fusion—let a shared-weight backbone exploit the two views far better. On the DvXray benchmark, the resulting DAGNet outperforms the previous state-of-the-art AHCR on every tested backbone, with test mAP gains of roughly 1.5 to 2.4 percentage points and gains up to 4.8 points over the dual-view baseline. The authors intend the design as a general upgrade path for dual-view security inspection, not a dataset-specific trick.

What carries the argument

The load-bearing machinery is a shared-weight backbone flanked by three modules: FDIM (creates a learnable, view-mixed frequency-domain filter via FFT, modulates the spectrum, and inverse-FFTs back), DVHEM (adds position encoding, swaps queries and keys between views in multi-head cross-attention, then produces hierarchical guidance maps through depthwise convolution and 7×7 convolutions), and CGFM (applies CBAM to each view, concatenates, normalizes, and merges through a depthwise convolution branch plus a 1×1 convolution branch). The argument is that each module handles one stage of the fusion pipeline—frequency-level enhancement, spatial and semantic alignment, and final fusion—so their combination closes the gap left by earlier methods.

What would settle it

A concrete check would be to train DAGNet and AHCR on DvXray with matched FLOPs or parameter counts, repeated over several random seeds, and compare test mAP distributions. If the gains shrink to within standard deviation—or if an equally large gain appears when AHCR is given the same FLOPs budget—the paper's central attribution to FDIM, DVHEM, and CGFM would be refuted.

Watch

Extended reading notes

Core claim

The central discovery is that inter-view information in dual-view X-ray images can be mined at three complementary levels: in the frequency domain, where FDIM adaptively amplifies useful bands and suppresses noisy ones using filters built jointly from both views; across scales, where DVHEM uses multi-head cross-attention plus hierarchical guidance to align features from the vertical and horizontal views; and at fusion time, where CGFM uses channel-spatial attention and a dual-branch bottleneck to combine the views while suppressing redundancy. On DvXray, these modules act synergistically: full DAGNet reaches 85.3% validation mAP versus 80.9% for the ResNet50 dual baseline, and the test mAP advantage over AHCR holds for ResNet50, ResNeXt50, RegNet, and ConvNeXt Tiny. The paper frames this as evidence that attention-guided, multi-level feature interaction solves the information-loss problem in dual-view fusion better than earlier single-level refinement.

Load-bearing premise

The load-bearing premise is that the reported mAP differences come from the proposed modules themselves rather than from the extra computation DAGNet uses or from random variation on a single data split.

Editorial extensions

If this is right

  • On the DvXray benchmark, DAGNet improves test mAP by 1.5–2.4 points over AHCR on all four backbones tested, and by 4.8 points over the ResNet50 dual-view baseline.
  • The modules are backbone-agnostic in the experiments: the same three-module architecture raises performance for ResNet50, ResNeXt50, RegNet x 3.2gf, and ConvNeXt Tiny.
  • Ablations indicate the modules are complementary but not uniformly additive: CGFM alone gives the largest single-module gain, while CGFM + FDIM performs worse than either alone, suggesting redundancy between frequency-domain and convolution-guided fusion under some conditions.
  • The authors formulate dual-view inspection as multi-label classification rather than geometric detection, so the method focuses on attribute-level feature associations and avoids explicit cross-view geometric matching.
  • Because the full model's gains are reported with higher FLOPs than the comparison method, the practical efficiency claim rests on the performance-per-computation trade-off being acceptable for security screening hardware.

Reading between the lines

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

  • Editorial extension: the frequency-domain interaction idea is inherently view-agnostic, so a natural test is whether the same module helps with three or more views by pooling filters across all view pairs rather than just two.
  • Editorial extension: because DAGNet's gains are reported on a single train/validation/test split with no repeated seeds, a capacity-matched and multi-seed comparison would clarify how much of the improvement is due to the modules rather than extra FLOPs.
  • Editorial extension: the same three-stage recipe—frequency interaction, cross-attention alignment, and guided fusion—could be applied to other dual-input tasks such as stereo depth estimation or RGB-depth classification, where misalignment and redundancy are also central difficulties.
  • Editorial extension: the paper evaluates image-level classification only; extending the framework to per-object detection heads would test whether the fused features preserve localization information, not just label presence.
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 proposes DAGNet, a dual-view X-ray security inspection network that combines three modules—FDIM (frequency-domain interaction), DVHEM (dual-view hierarchical enhancement), and CGFM (convolutional guided fusion)—on a shared-weight backbone. The task is formulated as multi-label, multi-category classification, and experiments on the DvXray dataset are reported for four backbone architectures (ResNet50, ResNeXt50, RegNet, ConvNeXt Tiny), comparing against a dual-view baseline and AHCR. The central claim is that DAGNet outperforms existing state-of-the-art approaches across these backbones, with code made publicly available.

Significance. If the reported results are reliable, DAGNet offers a practical and modular architecture for dual-view X-ray contraband classification, and the open-source implementation is a valuable contribution to a relatively underexplored application area. Testing across four diverse backbones is also a strength for generalization claims. However, the current evidence is not sufficient to establish the central state-of-the-art claim: the evaluation rests on a single split with no error bars, the proposed model consistently has higher FLOPs than the comparison method, and several specific numbers in the text contradict the tables. These issues are fixable with additional experiments and careful reporting, but they currently leave the causal attribution of the gains to the proposed modules unproven.

major comments (5)
  1. [Section IV-B, Table I] The state-of-the-art claim rests on a single 7:2:1 split with no repeated seeds, standard deviations, or significance testing. Reported margins over AHCR are small (e.g., ResNet50 test mAP 85.3 vs 83.8; ConvNeXt Tiny 91.0 vs 89.3) and could be within run-to-run variability for multi-label mAP. Please repeat experiments over at least three seeds and report mean and standard deviation, or provide a statistical significance test.
  2. [Table I, Section III] DAGNet's FLOPs are higher than AHCR's in every backbone configuration (ResNet50: 15.671G vs 12.409G; ConvNeXt Tiny: 12.356G vs 11.891G), and parameters also increase in most cases. The reported gains therefore cannot be attributed to the proposed modules rather than added capacity. Please add capacity-matched baselines (e.g., AHCR or the dual-view baseline augmented with additional layers to match DAGNet's FLOPs/parameters) or otherwise control for computational cost.
  3. [Section IV-D, Table I] The text states that DAGNet achieves 6.4%, 1.8%, 1.9%, and 2.5% test mAP improvements over the four baselines, but Table I shows absolute differences of 4.8, 2.9, 4.3, and 2.5 percentage points, respectively. Please clarify whether these are absolute or relative improvements and correct the text so that the numbers match the table.
  4. [Section IV-E, Table II] The ablation prose says that 'the combination of CGFM and FDIM performed slightly worse than using the FDIM module alone,' but Table II reports CGFM+FDIM at 83.3% val mAP versus FDIM alone at 81.4%, i.e., the combination performs better. This contradiction undermines the discussion of module redundancy and synergy and must be reconciled.
  5. [Section II, Section IV-D] The abstract and introduction claim that DAGNet outperforms 'existing state-of-the-art approaches,' but the experiments compare only against AHCR among dual-view methods, while other cited dual-view or transformer-based methods (e.g., [15]) are not evaluated. If those methods are not directly comparable or are not publicly available, please state this explicitly; otherwise the SOTA claim is broader than the evidence.
minor comments (6)
  1. [Abstract] There are grammar issues: 'scanner is' should be 'scanners are' and 'they struggles' should be 'they struggle.'
  2. [Section IV-B, Eq. (17)] The text says MultiLabelSoftMarginLoss is suitable for cases where labels are mutually exclusive, but this loss is designed for multi-label, non-exclusive settings; the description should be corrected.
  3. [Section III-B, Eqs. (1) and (6)] The notation is confusing: the input to the inverse FFT in Eq. (6) should be the filtered spectrum, but the expression 'F_X' · D_X' mixes the frequency-domain feature F_X' with a filter D_X whose dimensions (H x floor(W/2)) are not clearly compatible with the full complex spectrum; please clarify the shapes and the exact filtering operation.
  4. [References] Reference [12] lists placeholder authors 'F. Author, S. Author, and T. Author,' and reference [19] contains 'xxxxx' as page numbers; these should be completed or replaced with correct citations.
  5. [Section I and Section III-A] Figures 1 and 2 both appear with the caption 'Overall architecture of the proposed framework,' which is likely a typographical duplication; the captions should be distinguished.
  6. [Section IV-D] There is a typographical error in 'both the OL and SD)views': an unmatched parenthesis appears before 'views.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: DAGNet is an empirical architecture paper evaluated on an external held-out benchmark.

full rationale

DAGNet's central claim is that its proposed modules (FDIM, DVHEM, CGFM) improve mAP on the externally maintained DvXray dataset. There is no derivation chain in which a target quantity is defined in terms of the prediction or in which fitted parameters are renamed as predictions. Table I reports test mAP on a held-out 10% split of DvXray; the model is trained with a standard multi-label loss (Eq. 17) and evaluated with standard mAP (Eq. 22), so the reported numbers are empirical outcomes rather than identities. The paper does not invoke any uniqueness theorem, and no load-bearing argument reduces to a self-citation: the reference list contains no works by the present authors, and the main comparison baseline (AHCR) is an independent external method. Architecture selection on the validation split is ordinary model development, not circularity. The weaknesses noted by a skeptical reader, such as missing error bars, single-split evaluation, and higher FLOPs than AHCR, concern causal attribution and statistical confidence; they are correctness-risk issues and do not make the evaluation circular. Under the given standards, the correct finding is no significant circularity.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The ledger is dominated by data and task assumptions plus hand-chosen hyperparameters, because the central claim is empirical. No new physical entities are introduced; the modules proposed (FDIM, DVHEM, CGFM) are architectural components evaluated within the paper, not standalone entities with independent falsifiable handles, so the invented_entities list is empty.

free parameters (7)
  • Base learning rate and weight decay
    AdamW is named but the learning rate and weight decay values are absent, so the optimizer setup is not fully specified.
  • Random seed and split assignment
    The 7:2:1 split lacks a seed or stratification description; all numbers are single-run.
  • EMA decay factor
    EMA is used but its decay parameter is not reported.
  • DVHEM head count and MLP widths
    Multi-head cross-attention is described generically; head count and hidden widths are left to the code.
  • FDIM filter initialization and scaling
    The learnable frequency filter K is H x floor(W/2), but initialization and adaptive scaling of the complex filter are unspecified.
  • Epochs, batch size, and workers = 60 epochs, batch size 64, 32 workers
    These fixed choices affect convergence and throughput; they are reasonable but arbitrary.
  • Loss function = MultiLabelSoftMarginLoss
    This loss choice affects the optimization surface; the formula reduces to per-label BCE.
assumptions (6)
  • domain assumption DvXray image-level annotations are accurate and the dataset is reliable.
    All results depend on this external dataset; the paper states annotations were verified by security personnel, but the current authors did not audit them.
  • domain assumption The 7:2:1 split is random and prevents leakage between the two views of the same scanned bag.
    The splitting procedure is not described; if paired views are split across train and test sets, the test mAP would be inflated.
  • domain assumption ImageNet-1K pretrained weights transfer to X-ray security imagery.
    Every backbone is initialized with ImageNet-1K weights; the paper provides no evidence that this initialization is appropriate for X-ray images.
  • domain assumption Multi-label image-level classification is an appropriate formulation for the security inspection task.
    The model does not localize objects; the paper asserts this is a mainstream solution, but it does not compare with detection-based approaches.
  • ad hoc to paper A learnable complex filter with shape H x floor(W/2) in the Fourier domain is a suitable inductive bias for X-ray feature enhancement.
    FDIM assumes contraband-related features are frequency-separable in this specific way; no analysis or visualization supports this assumption.
  • domain assumption Shared weights in the backbone produce comparable feature spaces for the two views.
    Cross-attention between views requires aligned semantic features; shared weights encourage but do not guarantee this alignment.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DAGNet: A Dual-View Attention-Guided Network for Efficient X-ray Security Inspection." pith.science (2026). https://pith.science/paper/B4O2G7DL

@misc{pith2026250201710,
  author       = {Pith},
  title        = {Pith review of: DAGNet: A Dual-View Attention-Guided Network for Efficient X-ray Security Inspection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B4O2G7DL}},
  note         = {Machine review of arXiv:2502.01710}
}
read the original abstract

With the rapid development of modern transportation systems and the exponential growth of logistics volumes, intelligent X-ray-based security inspection systems play a crucial role in public safety. Although single-view X-ray baggage scanner is widely deployed, they struggles to accurately identify contraband in complex stacking scenarios due to strong viewpoint dependency and inadequate feature representation. To address this, we propose a Dual-View Attention-Guided Network for Efficient X-ray Security Inspection (DAGNet). This study builds on a shared-weight backbone network as the foundation and constructs three key modules that work together: (1) Frequency Domain Interaction Module (FDIM) dynamically enhances features by adjusting frequency components based on inter-view relationships; (2) Dual-View Hierarchical Enhancement Module (DVHEM) employs cross-attention to align features between views and capture hierarchical associations; (3) Convolutional Guided Fusion Module (CGFM) fuses features to suppress redundancy while retaining critical discriminative information. Collectively, these modules substantially improve the performance of dual-view X-ray security inspection. Experimental results demonstrate that DAGNet outperforms existing state-of-the-art approaches across multiple backbone architectures. The code is available at:https://github.com/ShilongHong/DAGNet.

Figures

Figures reproduced from arXiv: 2502.01710 by the authors.

Figure 1
Figure 1. Overall architecture of the proposed framework. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overall architecture of the proposed framework. [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The overall architecture of the proposed Frequency Domain Interaction [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Overall architecture of the Dual-View Hierarchical Enhancement [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: The overall architecture of the proposed Convolutional Attention [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages

  1. [15]

    Transformer-based dual-view x-ray security inspection image analysis,

    X. Meng, H. Feng, Y . Ren, H. Zhang, W. Zou, and X. Ouyang, “Transformer-based dual-view x-ray security inspection image analysis,” Engineering Applications of Artificial Intelligence , vol. 138, p. 109382, 2024

    work page 2024

  2. [1]

    An evaluation of region based object detection strategies within x-ray baggage security imagery,

    S. Akcay and T. P. Breckon, “An evaluation of region based object detection strategies within x-ray baggage security imagery,” in 2017 IEEE International Conference on Image Processing (ICIP) . IEEE, 2017, pp. 1337–1341

    work page 2017

  3. [2]

    Towards real-world prohibited item detection: A large-scale x-ray benchmark,

    B. Wang, L. Zhang, L. Wen, X. Liu, and Y . Wu, “Towards real-world prohibited item detection: A large-scale x-ray benchmark,” in 2021 IEEE/CVF International Conference on Computer Vision (ICCV) , 2021, pp. 5392–5401

    work page 2021

  4. [3]

    Detecting overlapped objects in x-ray security imagery by a label-aware mechanism,

    C. Zhao, L. Zhu, S. Dou, W. Deng, and L. Wang, “Detecting overlapped objects in x-ray security imagery by a label-aware mechanism,” IEEE Transactions on Information Forensics and Security , vol. 17, pp. 998– 1009, 2022

    work page 2022

  5. [4]

    Using deep convolutional neural network architectures for object clas- sification and detection within x-ray baggage security imagery,

    S. Akc ¸ay, M. E. Kundegorski, C. G. Willcocks, and T. P. Breckon, “Using deep convolutional neural network architectures for object clas- sification and detection within x-ray baggage security imagery,” IEEE Trans. Inf. Forensics Security, vol. 13, no. 9, pp. 2203–2215, Sep. 2018

    work page 2018

  6. [5]

    Sixray: A large-scale security inspection x-ray benchmark for prohibited item discovery in overlapping images,

    C. Miao and al., “Sixray: A large-scale security inspection x-ray benchmark for prohibited item discovery in overlapping images,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR) , Jun. 2019, pp. 2119–2128

    work page 2019

  7. [6]

    Towards real-world x-ray security inspection: A high- quality benchmark and lateral inhibition module for prohibited items detection,

    R. Tao and al., “Towards real-world x-ray security inspection: A high- quality benchmark and lateral inhibition module for prohibited items detection,” in Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV) , Oct. 2021, pp. 10 923–10 932

    work page 2021

  8. [7]

    Towards dual- view x-ray baggage inspection: A large-scale benchmark and adaptive hierarchical cross refinement for prohibited item discovery,

    B. Ma, T. Jia, M. Li, S. Wu, H. Wang, and D. Chen, “Towards dual- view x-ray baggage inspection: A large-scale benchmark and adaptive hierarchical cross refinement for prohibited item discovery,”IEEE Trans- actions on Information Forensics and Security , 2024

    work page 2024

Show all 35 references
  1. [8]

    Advancements in machine learning techniques for threat item detection in x-ray images: a comprehensive survey,

    A. Singh and Dhiraj, “Advancements in machine learning techniques for threat item detection in x-ray images: a comprehensive survey,” International Journal of Multimedia Information Retrieval , vol. 13, no. 4, p. 40, 2024

    2024

  2. [9]

    Occluded prohibited items detection: An x-ray security inspection benchmark and de-occlusion attention module,

    Y . Wei, R. Tao, Z. Wu, Y . Ma, L. Zhang, and X. Liu, “Occluded prohibited items detection: An x-ray security inspection benchmark and de-occlusion attention module,” in Proceedings of the 28th ACM international conference on multimedia , 2020, pp. 138–146

    2020

  3. [10]

    Cfpa-net: Cross-layer feature fusion and parallel attention network for detection and classification of prohib- ited items in x-ray baggage images,

    Y . Wei, Y . Wang, and H. Song, “Cfpa-net: Cross-layer feature fusion and parallel attention network for detection and classification of prohib- ited items in x-ray baggage images,” in 2021 IEEE 7th International Conference on Cloud Computing and Intelligent Systems (CCIS) , 2...

    2021

  4. [11]

    A defect recognition method for low-quality weld image based on consistent multiscale fea- ture mapping,

    J. Liu, X. Liu, F. Qu, H. Zhang, and L. Zhang, “A defect recognition method for low-quality weld image based on consistent multiscale fea- ture mapping,” IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1–11, 2022

    2022

  5. [12]

    Re-bifpn: Refining bifpn for accurate object detection in images,

    F. Author, S. Author, and T. Author, “Re-bifpn: Refining bifpn for accurate object detection in images,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 1234–1245

    2022

  6. [13]

    Fdtnet: Enhanc- ing frequency-aware representation for prohibited object detection from x-ray images via dual-stream transformers,

    Z. Zhu, Y . Zhu, H. Wang, N. Wang, J. Ye, and X. Ling, “Fdtnet: Enhanc- ing frequency-aware representation for prohibited object detection from x-ray images via dual-stream transformers,” Engineering Applications of Artificial Intelligence , vol. 133, p. 108076, 2024

    2024

  7. [14]

    Dual-view x-ray detection: Can ai detect prohibited items from dual-view x-ray images like humans?

    R. Tao, H. Wang, Y . Guo, H. Chen, L. Zhang, X. Liu, Y . Wei, and Y . Zhao, “Dual-view x-ray detection: Can ai detect prohibited items from dual-view x-ray images like humans?” arXiv preprint arXiv:2411.18082, 2024

    2024 arXiv

  8. [16]

    High performance depthwise and pointwise convolutions on mobile devices,

    P. Zhang, E. Lo, and B. Lu, “High performance depthwise and pointwise convolutions on mobile devices,” inProceedings of the AAAI Conference on Artificial Intelligence , vol. 34, no. 04, 2020, pp. 6795–6802

    2020

  9. [17]

    R. C. Gonzalez and R. E. Woods, Digital Image Processing , 4th ed. Pearson, 2018

    2018

  10. [18]

    Fft-based dynamic token mixer for vision,

    Y . Tatsunami and M. Taki, “Fft-based dynamic token mixer for vision,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 14, 2024, pp. 15 328–15 336

    2024

  11. [19]

    Guided real image dehazing using ycbcr color space,

    W. Fang, J. Fan, Y . Zheng, J. Weng, Y . Tai, and J. Li, “Guided real image dehazing using ycbcr color space,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025, pp. xxxxx–xxxxx

    2025

  12. [20]

    Image transformers,

    A. Dosovitskiy et al., “Image transformers,” in Proceedings of the 37th International Conference on Machine Learning (ICML) , 2020, pp. 298– 306

    2020

  13. [21]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, H. Li, B. Guo, and L. Zhang, “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV) , 2021, pp. 10 012–10 022

    2021

  14. [22]

    Mobilenets: Efficient convolutional neural networks for mobile vision applications,

    A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam, “Mobilenets: Efficient convolutional neural networks for mobile vision applications,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, p...

    2017

  15. [23]

    Xception: Deep learning with depthwise separable convolu- tions,

    F. Chollet, “Xception: Deep learning with depthwise separable convolu- tions,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1251–1258

    2017

  16. [24]

    Cbam: Convolutional block attention module,

    S. Woo, J. Park, J.-Y . Lee, and I. S. Kweon, “Cbam: Convolutional block attention module,” in Proceedings of the European Conference on Computer Vision, 2018, pp. 3–19

    2018

  17. [25]

    Cspnet: A new backbone that can enhance learning capability of cnn,

    C.-Y . Wang, H.-Y . Mark Liao, Y .-H. Wu, P.-Y . Chen, J.-W. Hsieh, and I.- H. Yeh, “Cspnet: A new backbone that can enhance learning capability of cnn,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops , 2020, pp. 390–391

    2020

  18. [26]

    Multi-view x-ray r-cnn,

    J.-M. O. Steitz, F. Saeedan, and S. Roth, “Multi-view x-ray r-cnn,” in German Conference on Pattern Recognition . Springer, 2018, pp. 153– 168

    2018

  19. [27]

    Multi-view object detection using epipolar constraints within cluttered x-ray security imagery,

    B. K. Isaac-Medina, C. G. Willcocks, and T. P. Breckon, “Multi-view object detection using epipolar constraints within cluttered x-ray security imagery,” inProceedings of the 25th International Conference on Pattern Recognition (ICPR), 2021, pp. 9889–9896

    2021

  20. [28]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, S. Chintala, G. Chanan, E. Yang, Z. DeVito, M. Killeen, Y . Lin, A. Nath, A. Lerer et al. , “Pytorch: An imperative style, high-performance deep learning library,” Proceedings of NeurIPS, vol. 32, 2019

    2019

  21. [29]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” in International Conference on Learning Representations (ICLR) , 2019. [Online]. Available: https://arxiv.org/abs/1711.05101

    2019 arXiv

  22. [30]

    Mixed precision training,

    P. Micikevicius, S. Narang, G. Alben, G. Diamos, E. Elsen, B. Ginsburg et al. , “Mixed precision training,” in Proceedings of the International Conference on Learning Representations (ICLR) , 2018

    2018

  23. [31]

    Acceleration of stochastic approximation by averaging,

    B. T. Polyak and A. Juditsky, “Acceleration of stochastic approximation by averaging,” in SIAM Journal on Control and Optimization , vol. 30, no. 4, 1992, pp. 838–855

    1992

  24. [32]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2016, pp. 770–778

    2016

  25. [33]

    Aggregated residual transformations for deep neural networks,

    S. Xie, R. Girshick, P. Doll ´ar, Z. Tu, and K. He, “Aggregated residual transformations for deep neural networks,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2017, pp. 1492–1500

    2017

  26. [34]

    De- signing network design spaces,

    I. Radosavovic, R. P. Kosaraju, R. Girshick, K. He, and P. Doll ´ar, “De- signing network design spaces,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2020, pp. 10 428–10 436

    2020

  27. [35]

    A convnet for the 2020s,

    Z. Liu, H. Mao, C.-Y . Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A convnet for the 2020s,” arXiv preprint arXiv:2201.03545 , 2022

    2022 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.