Pith. sign in

REVIEW 3 major objections 5 minor 63 references

Efficient Detection Framework Adaptation for Edge Computing: A Plug-and-play Neural Network Toolbox Enabling Edge Deployment

T0 review · 3 major / 5 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 A plug-and-play toolbox of three modules shrinks CNN object detectors for edge hardware while raising accuracy, demonstrated on a new helmet-band safety dataset.

desk verdict A useful engineering toolbox paper with a new safety dataset, but the complexity proof doesn't match the architecture and the HBDD real-world validation is weakened by synthetic negative samples. read the letter →

arxiv 2412.18230 v1 pith:ACUKMGI3 submitted 2024-12-24 cs.CV

Jiaqi Wu , Shihao Zhang , Simin Chen , Lixu Wang , Zehua Wang , Wei Chen , Fangyuan He , Zijian Tian
show 2 more authors
F. Richard Yu Victor C. M. Leung
This is my paper · ORCID
classification cs.CV
keywords edgecomputingobjectdetectionplug-and-playmodulesreparameterizationlightweightneuralnetworkscross-attentionhelmetbandYOLO
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 settle a practical trade-off: edge deployment of object detectors should not force a big accuracy drop. It proposes ED-TOOLBOX, a set of plug-and-play components that replace or insert into standard CNN detector parts (Backbone, Neck, Head), and reports that the resulting ED-YOLO reaches 91.34 mAP on a new helmet-band dataset with 10.9 million parameters and 19.7 GFLOPs, cutting parameters by 17% and FLOPs by 23% relative to YOLOv8-s. The authors also build HBDD, a dataset covering unfastened helmet chin straps, a safety-relevant class that prior helmet detectors ignore. If the claim holds, time-sensitive surveillance on low-power devices can run accurate small-object detection without sending video to the cloud. The authors explicitly limit the toolbox to CNN-based detectors, leaving Transformer detectors and tasks like segmentation and tracking outside its current scope.

What carries the argument

Three named components carry the argument. Rep-DConvNet (Reparameterized Dynamic Convolutional Network) is a lightweight block that splits channels into groups, applies weighted horizontal ($1\times3$), vertical ($3\times1$), and square ($3\times3$) depthwise convolutions with a residual, then channel-shuffles; during inference the branches fuse into one convolution, giving multi-branch training with lower inference cost. SC-A (Sparse Cross-Attention) is a parameter-free attention inside the Joint Module: it pools feature maps along horizontal and vertical directions to form $H+W$ tokens, computes cross-attention between current-module and upstream-module tokens, then expands back to spatial maps, capturing long-range dependencies at $\mathcal{O}(H+W)$ complexity instead of Non-local's $\mathcal{O}(H\cdot W \cdot h \cdot w)$. The Efficient Head replaces YOLOv8's dual-branch decoupled head with a single-branch Ghost-convolution head and adds a $160\times160$ detection branch for small objects.

What would settle it

Run ED-YOLO trained only on HBDD on a held-out set of genuine, unedited surveillance clips from working mines showing unfastened chin straps, and compare its mAP with the reported 91.34; a large drop would show the simulation-generated negative samples are not faithful proxies for real deployment.

Watch

Extended reading notes

Core claim

The central claim is that detection-specific plug-and-play modules are what let a detector stay accurate after being compressed for the edge, and the paper demonstrates this with three such modules. Rep-DConvNet replaces standard convolutions with weighted horizontal, vertical, and square depthwise branches plus a residual, fusing the branches at inference time; on a large public classification benchmark it uses 1.8 million parameters and 154.9 million FLOPs at 72.7% accuracy, below most lightweight baselines in size. SC-A, the parameter-free Sparse Cross-Attention inside the Joint Module, pools feature maps into horizontal and vertical tokens, computes cross-attention between the current module and its upstream neighbor, and expands the result back, capturing long-range dependencies without added parameters. The Efficient Head replaces YOLOv8's dual-branch decoupled head with a single-branch Ghost-convolution head and adds a $160\times160$ detection branch aimed at small objects. On the new HBDD benchmark, ED-YOLO reaches 91.34 mAP with 10.9 million parameters and 19.7 GFLOPs, and in a surveillance-system simulation on a low-power embedded GPU it sustains 21.6 FPS with 16.86 total FPS including data transmission, an accuracy level close to much heavier cloud models.

Load-bearing premise

The new dataset's negative examples (helmets worn without fastened straps) are partly simulation-generated, and the paper assumes these synthetic images match real mine-surveillance footage closely enough that the reported accuracy survives deployment.

Editorial extensions

If this is right

  • ED-YOLO trained with the toolbox can be deployed on a low-power embedded GPU and still detect small helmet straps in near-real time, with total throughput of 16.86 FPS including data transmission in the simulated surveillance system.
  • The same Rep-DConvNet backbone inserted into SSD lifts mAP by 9.22 points while cutting FLOPs by 16%, indicating the components transfer beyond YOLO to at least one other CNN detector.
  • Removing any single component degrades the result: without Rep-DConvNet mAP falls 6.19 points and FLOPs rise 74.9%, and without the Joint Module mAP falls 3.61 points.
  • The HBDD dataset introduces a safety-critical class absent from prior helmet datasets, unfastened chin straps, and the paper shows the toolbox detects these small targets when standard lightweight models miss them.
  • The authors explicitly limit the toolbox to CNN-based detectors; Transformer detectors and other tasks such as segmentation and tracking are not supported.

Reading between the lines

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

  • If HBDD's simulated negative images are visually faithful to real mine footage, the reported accuracy should transfer; the quickest check is a test set of genuine unfastened-strap surveillance images, which the authors do not provide.
  • The SC-A tokenization (pooling to $H+W$ tokens) is a general complexity reduction that could be applied to other pairwise attention modules, such as multi-scale feature fusion for segmentation, although the paper does not claim this.
  • The combination of a parameter-free attention module with reparameterized convolutions suggests that both training-time structural augmentation and inference-time fusion contribute to the small-object gain; an ablation varying only the fusion without the weighted branches could separate the two effects.
  • The authors' future plan to fine-tune a pre-trained general module per task points toward a foundation-model-style deployment kit, but nothing in the current experiments tests that path.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes ED-TOOLBOX, a set of plug-and-play modules for adapting CNN-based object detectors to edge devices: Rep-DConvNet (a reparameterized multi-branch backbone block), SC-A (a sparse cross-attention module forming a Joint Module), and an Efficient Head for YOLO-style detectors. The authors introduce the Helmet Band Detection Dataset (HBDD) for detecting improper helmet wearing, including chin-strap fastening, and evaluate ED-YOLO and ED-SSD on ImageNet, COCO, HBDD, and a Jetson TX2-based IoVT surveillance simulation. The central empirical claim is that ED-YOLO achieves 91.34 mAP on HBDD with 10.9M parameters and 19.7 GFLOPs, reducing parameters by 17% and FLOPs by 23% versus YOLOv8-s while improving mAP, and that it reaches 21.6 FPS on the edge device.

Significance. If the empirical results hold, the paper provides a practical and modular recipe for shrinking YOLO-family detectors while preserving or improving accuracy, plus a new dataset that targets an underexplored safety task (chin-strap detection). The zero-parameter SC-A module, the component-wise ablation, and the edge simulation are valuable assets. However, the theoretical complexity proofs contain a mismatch with the described architecture, and the real-world validity of the HBDD evaluation is weakened by the use of simulation-generated negative samples. The COCO evidence for generality is reported only in prose, not in a table. The core empirical claim is plausible but needs these points addressed before publication.

major comments (3)
  1. [Section III-B and IV-B] Eq. (2) defines branches with 1×3, 3×1, and 3×3 kernels but uses the sgn function in a way that does not select a single branch: with the standard convention sgn(0)=0, group i=1 receives negative contributions from the other branches; with sgn(0)=1, group i=2 receives two branches. Please replace this with an explicit per-group branch assignment. Moreover, Assumption 1 in Section IV-B states that all three convolution branches use 3×3 kernels, contradicting Eq. (2), and the complexity counts in Eqs. (10)–(15) use 9 multiply-adds for every branch. Since Theorem 1 is the basis for the 'lower complexity than RepVGG' claim, the assumption must be aligned with the actual kernels (1×3 and 3×1 each cost 3 multiply-adds) or the proof revised.
  2. [Section V-A and V-C] The HBDD negative samples are supplemented by simulation-generated images because real surveillance rarely contains improperly worn helmets, and the 8:2 random split is over the combined set, so the test set likely contains synthetic negatives. The paper gives no per-class AP, no separate results on real-only images, and no domain-gap analysis. As the practical-impact claim depends on accurate hatband detection in real CCTV footage, the reported mAP (Table V: 91.34; Table VII: 90.21) may not transfer to deployment. Please report per-class AP (especially 'noband'), evaluate on a real-only subset if feasible, or explicitly state that the test set includes synthetic images and temper the 'real-world validation' claim accordingly.
  3. [Section V-C, 'Numerical Results'] The COCO2017 comparison is given only in prose (ED-YOLO 61.5% vs YOLOv8-s 57.9%) without a table, training schedule, input resolution, or the other methods' COCO numbers. This result is load-bearing for the generality claim (RQ4). Please provide a full COCO comparison table with the same details as the HBDD experiments, or clearly label these numbers as preliminary.
minor comments (5)
  1. [Section V-C] 'Compared to the original ED-SSD' should read 'compared to the original SSD-VGG' (the same typo appears in the following sentence).
  2. [Section V-C] The stated parameter reduction of '17%' for ED-YOLO is inconsistent with Table V, which gives 14.3M → 10.9M, a 23.8% reduction.
  3. [Section V-D] The ablation text says removing Rep-DConvNet 'increases FLOPs by 74.9%', but Table VI shows 28.519 vs 19.725 GFLOPs, a 44.6% increase.
  4. [Eq. (12)] The expression '= 9+4 / (27·Cin+1)' should be written as '(9+4)/(27·Cin+1)' to avoid ambiguity.
  5. [Throughout] Typos such as 'Adittionally', 'inferece', 'Deteails', and 'A Edge Detection Toolbox' should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed gains are tested against external benchmarks and a new dataset, and the complexity bounds follow from stated assumptions rather than from the results they purport to explain.

full rationale

The paper's central claims are empirical: ED-TOOLBOX components improve accuracy and reduce model scale on ImageNet, COCO-2017, and the newly introduced HBDD. These evaluations are self-contained against external benchmarks, so the reported mAP, parameter, and FLOPs numbers are not constructed from the test data. The theoretical complexity claims (Theorems 1 and 2) are derived from explicitly stated assumptions (e.g., equal input/output channels, approximately square inputs) and are algebraic comparisons against RepVGG and Non-local baselines; they do not presuppose the empirical accuracy results. The ablation study (Table VI) tests each component by removal, which is a standard causal attribution rather than a circular definition of the component's contribution. The only self-citations, such as Refs. [4], [6], and [12], appear as related-work context and future-work inspiration, and none is load-bearing for the paper's central derivation or evaluation. The use of simulation-generated images to supplement HBDD negative samples is a limitation for real-world transferability, but it is not circular: the model is trained and evaluated on a held-out split of that dataset, and the claim is about performance on the constructed benchmark rather than an assumption of the benchmark's outcome. Overall, no derivation step reduces, by construction or by self-citation, to its own inputs.

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

The theoretical complexity analysis rests on five stated assumptions, one of which (Assumption 1) is inconsistent with the actual network architecture. No new physical entities or fitted physical constants are introduced; the learnable weights in Rep-DConvNet are standard network parameters.

assumptions (5)
  • ad hoc to paper Assumption 1: The multi-branch network consists of four branches, each employing a 3x3 convolutional layer.
    Used in Theorem 1 to compute complexity, but contradicts the actual Rep-DConvNet branches (1x3, 3x1, 3x3) in Eq. (2).
  • domain assumption Assumption 2: Input and output channels are equal (Cin = Cout).
    Used to simplify the complexity ratio; holds only for the Basic Network, not for downsampling layers.
  • domain assumption Assumption 3: Channel shuffle cost c is negligible.
    Used to drop terms in the complexity ratio; plausible but not quantified.
  • domain assumption Assumption 4: Feature maps A and B have identical dimensions (H=h, W=w).
    Used in Theorem 2; not always true for neck connections between different scales in YOLO.
  • domain assumption Assumption 5: Input aspect ratio is 1.
    Used in Theorem 2 to simplify; detection inputs are often resized to square, so acceptable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Detection Framework Adaptation for Edge Computing: A Plug-and-play Neural Network Toolbox Enabling Edge Deployment." pith.science (2026). https://pith.science/paper/ACUKMGI3

@misc{pith2026241218230,
  author       = {Pith},
  title        = {Pith review of: Efficient Detection Framework Adaptation for Edge Computing: A Plug-and-play Neural Network Toolbox Enabling Edge Deployment},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ACUKMGI3}},
  note         = {Machine review of arXiv:2412.18230}
}
read the original abstract

Edge computing has emerged as a key paradigm for deploying deep learning-based object detection in time-sensitive scenarios. However, existing edge detection methods face challenges: 1) difficulty balancing detection precision with lightweight models, 2) limited adaptability of generalized deployment designs, and 3) insufficient real-world validation. To address these issues, we propose the Edge Detection Toolbox (ED-TOOLBOX), which utilizes generalizable plug-and-play components to adapt object detection models for edge environments. Specifically, we introduce a lightweight Reparameterized Dynamic Convolutional Network (Rep-DConvNet) featuring weighted multi-shape convolutional branches to enhance detection performance. Additionally, we design a Sparse Cross-Attention (SC-A) network with a localized-mapping-assisted self-attention mechanism, enabling a well-crafted joint module for adaptive feature transfer. For real-world applications, we incorporate an Efficient Head into the YOLO framework to accelerate edge model optimization. To demonstrate practical impact, we identify a gap in helmet detection -- overlooking band fastening, a critical safety factor -- and create the Helmet Band Detection Dataset (HBDD). Using ED-TOOLBOX-optimized models, we address this real-world task. Extensive experiments validate the effectiveness of ED-TOOLBOX, with edge detection models outperforming six state-of-the-art methods in visual surveillance simulations, achieving real-time and accurate performance. These results highlight ED-TOOLBOX as a superior solution for edge object detection.

Figures

Figures reproduced from arXiv: 2412.18230 by the authors.

Figure 1
Figure 1. The video surveillance system is based on an Internet of Video Things (IoVT) [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Different helmet-wearing behaviours. Red boxes indicate [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of the ED-TOOLBOX. It includes the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (12 more)
Figure 4
Figure 4. Figure 4: The structure of Rep-DConvNet. It decouples model [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Procedure of Sparse Cross-Attention (SC-A). At the [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 3
Figure 3. Figure 3: To further reduce the parameter count, we replace the [PITH_FULL_IMAGE:figures/full_fig_p007_3.png]
Figure 7
Figure 7. Figure 7: Assumption 3. In the Channel Shuffle, the parameter c is a small positive value, according to the low computational complexity characteristic of indexing, reordering, and any associated memory operations [27]. Theorem 1. Under Assumptions 1 to 3, the computational comp…
Figure 6
Figure 6. Figure 6: The example of HBDD. There are example of imges captured by real coal mine [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Structure of ED-YOLO. Based on the YOLO backbone model, YOLOv8-s, ED-TOOLBOX implements improvements for edge deployment. Specifically, Rep-DConvNet (as [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Comparison of detection results on our HBDD and the public COCO datasets. Lightweight models such as YOLOv7-tiny-SiLU, PPYOLOEs, and YOLOv5-s exhibit significant [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Compare the resource utilization of ED-YOLO with the n/s/m/l/x versions of the [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Comparison of feature extraction effects. Compared to the original DarkNet-53, our Backbone module (as shown in Fig. 7) [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Comparison of feature perception performance. According to Fig. 7, when the Neck does not incorporate the Joint Module, the model fails to capture small-sized objects, [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Comparison of detection results in the IoVT surveillance system. Compared to other lightweight models, such as YOLOv7-tiny-SiLU, PPYOLOEs, and YOLOv8-s, our [PITH_FULL_IMAGE:figures/full_fig_p014_12.png]
Figure 13
Figure 13. Figure 13: The delay comparison of two computing approaches. The delay difference is [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 40 canonical work pages

  1. [1]

    Yolov8-qsd: An improved small object detection algorithm for autonomous vehicles based on yolov8,

    H. Wang, C. Liu, Y . Cai, L. Chen, and Y . Li, “Yolov8-qsd: An improved small object detection algorithm for autonomous vehicles based on yolov8,” IEEE Transactions on Instrumentation and Measurement, 2024

    work page 2024

  2. [2]

    A generalist vision–language foundation model for diverse biomedical tasks,

    K. Zhang, R. Zhou, E. Adhikarla, Z. Yan, Y . Liu, J. Yu, Z. Liu, X. Chen, B. D. Davison, H. Ren et al., “A generalist vision–language foundation model for diverse biomedical tasks,” Nature Medicine, pp. 1–13, 2024

    2024

  3. [3]

    Transunet: Rethinking the u-net architecture design for medical image segmentation through the lens of transformers,

    J. Chen, J. Mei, X. Li, Y . Lu, Q. Yu, Q. Wei, X. Luo, Y . Xie, E. Adeli, Y . Wanget al., “Transunet: Rethinking the u-net architecture design for medical image segmentation through the lens of transformers,” Medical Image Analysis, vol. 97, p. 103280, 2024

    work page 2024

  4. [4]

    A lightweight small object detection method based on multi-layer coordination federated intelligence for coal mine iovt,

    J. Wu, R. Zheng, J. Jiang, Z. Tian, W. Chen, Z. Wang, F. R. Yu, and V . C. Leung, “A lightweight small object detection method based on multi-layer coordination federated intelligence for coal mine iovt,” IEEE Internet of Things Journal , 2024

    work page 2024

  5. [5]

    Fusion-driven deep feature network for enhanced object detection and tracking in video surveillance systems,

    D. K. Jain, X. Zhao, C. Gan, P. K. Shukla, A. Jain, and S. Sharma, “Fusion-driven deep feature network for enhanced object detection and tracking in video surveillance systems,” Information Fusion , vol. 109, p. 102429, 2024

    work page 2024

  6. [6]

    Small insulator defects detection based on multi-scale feature interaction transformer for uav-assisted power iovt,

    J. Wu, R. Jing, Y . Bai, Z. Tian, C. Wei, S. Zhang, F. R. Yu, and V . C. Leung, “Small insulator defects detection based on multi-scale feature interaction transformer for uav-assisted power iovt,” IEEE Internet of Things Journal, 2024

    work page 2024

  7. [7]

    Edge comput- ing with artificial intelligence: A machine learning perspective,

    H. Hua, Y . Li, T. Wang, N. Dong, W. Li, and J. Cao, “Edge comput- ing with artificial intelligence: A machine learning perspective,” ACM Computing Surveys, vol. 55, no. 9, pp. 1–35, 2023

    2023

  8. [8]

    Edge-computing-driven internet of things: A survey,

    L. Kong, J. Tan, J. Huang, G. Chen, S. Wang, X. Jin, P. Zeng, M. Khan, and S. K. Das, “Edge-computing-driven internet of things: A survey,” ACM Computing Surveys , vol. 55, no. 8, pp. 1–41, 2022

    work page 2022

Show all 63 references
  1. [9]

    Cloud computing in construction industry: Use cases, benefits and challenges,

    S. A. Bello, L. O. Oyedele, O. O. Akinade, M. Bilal, J. M. D. Delgado, L. A. Akanbi, A. O. Ajayi, and H. A. Owolabi, “Cloud computing in construction industry: Use cases, benefits and challenges,” Automation in Construction, vol. 122, p. 103441, 2021

    2021

  2. [10]

    Edge yolo: Real-time intelligent object detection system based on edge-cloud cooperation in autonomous vehicles,

    S. Liang, H. Wu, L. Zhen, Q. Hua, S. Garg, G. Kaddoum, M. M. Hassan, and K. Yu, “Edge yolo: Real-time intelligent object detection system based on edge-cloud cooperation in autonomous vehicles,” IEEE Transactions on Intelligent Transportation Systems , vol. 23, no. 12, pp. 25 ...

    2022

  3. [11]

    Cooperative multi- camera vehicle tracking and traffic surveillance with edge artificial intelligence and representation learning,

    H. F. Yang, J. Cai, C. Liu, R. Ke, and Y . Wang, “Cooperative multi- camera vehicle tracking and traffic surveillance with edge artificial intelligence and representation learning,” Transportation research part C: emerging technologies , vol. 148, p. 103982, 2023

    2023

  4. [12]

    A small object real-time detection method for power line inspection in low-illuminance environments,

    Y . Zhao, J. Wu, W. Chen, Z. Wang, Z. Tian, F. R. Yu, and V . C. Leung, “A small object real-time detection method for power line inspection in low-illuminance environments,” IEEE Transactions on Emerging Topics in Computational Intelligence , 2024. IEEE TRANSACTIONS ON MOBILE...

    2024

  5. [13]

    Mobilenets: Efficient convolutional neural networks for mobile vision applications,

    A. G. Howard, “Mobilenets: Efficient convolutional neural networks for mobile vision applications,” arXiv preprint arXiv:1704.04861 , 2017

    2017 arXiv

  6. [14]

    Mobilevit: Light-weight, general-purpose, and mobile-friendly vision transformer,

    S. Mehta and M. Rastegari, “Mobilevit: Light-weight, general-purpose, and mobile-friendly vision transformer,” in International Conference on Learning Representations, 2022

    2022

  7. [15]

    Efficientformer: Vision transformers at mobilenet speed,

    Y . Li, G. Yuan, Y . Wen, J. Hu, G. Evangelidis, S. Tulyakov, Y . Wang, and J. Ren, “Efficientformer: Vision transformers at mobilenet speed,” Ad- vances in Neural Information Processing Systems , vol. 35, pp. 12 934– 12 949, 2022

    2022

  8. [16]

    The yolo framework: A comprehensive review of evolution, applications, and benchmarks in object detection,

    M. L. Ali and Z. Zhang, “The yolo framework: A comprehensive review of evolution, applications, and benchmarks in object detection,” 2024

    2024

  9. [17]

    Ssd: Single shot multibox detector,

    W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C.-Y . Fu, and A. C. Berg, “Ssd: Single shot multibox detector,” in Computer Vision– ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14 . Springer, 2016, pp. 21–37

    2016

  10. [18]

    Real- time automatic helmet detection of motorcyclists in urban traffic using improved yolov5 detector,

    W. Jia, S. Xu, Z. Liang, Y . Zhao, H. Min, S. Li, and Y . Yu, “Real- time automatic helmet detection of motorcyclists in urban traffic using improved yolov5 detector,” IET Image Processing

  11. [19]

    Deep learning based protective equipment detection on offshore drilling platform,

    F. Gong, X. Ji, W. Gong, and C. Gong, “Deep learning based protective equipment detection on offshore drilling platform,” Symmetry, 2021

    2021

  12. [20]

    Internet of video things: Next-generation iot with visual sensors,

    C. W. Chen, “Internet of video things: Next-generation iot with visual sensors,” IEEE Internet of Things Journal , vol. 7, no. 8, pp. 6676–6685, 2020

    2020

  13. [21]

    Automated exploration and implementation of distributed cnn inference at the edge,

    X. Guo, A. D. Pimentel, and T. Stefanov, “Automated exploration and implementation of distributed cnn inference at the edge,” IEEE Internet of Things Journal , vol. 10, no. 7, pp. 5843–5858, 2023

    2023

  14. [22]

    Roofsplit: an edge com- puting framework with heterogeneous nodes collaboration considering optimal cnn model splitting,

    Y . Huang, H. Zhang, X. Shao, X. Li, and H. Ji, “Roofsplit: an edge com- puting framework with heterogeneous nodes collaboration considering optimal cnn model splitting,” Future Generation Computer Systems, vol. 140, pp. 79–90, 2023

    2023

  15. [23]

    A comprehensive survey of neural architecture search: Challenges and solutions,

    P. Ren, Y . Xiao, X. Chang, P.-Y . Huang, Z. Li, X. Chen, and X. Wang, “A comprehensive survey of neural architecture search: Challenges and solutions,” ACM Computing Surveys (CSUR) , vol. 54, no. 4, pp. 1–34, 2021

    2021

  16. [24]

    Hadas: Hardware-aware dynamic neural architecture search for edge performance scaling,

    H. Bouzidi, M. Odema, H. Ouarnoughi, M. A. Al Faruque, and S. Niar, “Hadas: Hardware-aware dynamic neural architecture search for edge performance scaling,” in 2023 Design, Automation & Test in Europe Conference & Exhibition (DATE) . IEEE, 2023, pp. 1–6

    2023

  17. [25]

    Finch: Enhancing federated learning with hierarchical neural architecture search,

    J. Liu, J. Yan, H. Xu, Z. Wang, J. Huang, and Y . Xu, “Finch: Enhancing federated learning with hierarchical neural architecture search,” IEEE Transactions on Mobile Computing , 2023

    2023

  18. [26]

    Neural architecture search algo- rithm to optimize deep transformer model for fault detection in electrical power distribution systems,

    J. B. Thomas and K. Shihabudheen, “Neural architecture search algo- rithm to optimize deep transformer model for fault detection in electrical power distribution systems,” Engineering Applications of Artificial In- telligence, vol. 120, p. 105890, 2023

    2023

  19. [27]

    Shufflenet: An extremely effi- cient convolutional neural network for mobile devices,

    X. Zhang, X. Zhou, M. Lin, and J. Sun, “Shufflenet: An extremely effi- cient convolutional neural network for mobile devices,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 6848–6856

    2018

  20. [28]

    Repvgg: Mak- ing vgg-style convnets great again,

    X. Ding, X. Zhang, N. Ma, J. Han, G. Ding, and J. Sun, “Repvgg: Mak- ing vgg-style convnets great again,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 13 733–13 742

    2021

  21. [29]

    Pose guided anchoring for detecting proper use of personal protective equipment,

    R. Xiong and P. Tang, “Pose guided anchoring for detecting proper use of personal protective equipment,” Automation in Construction, vol. 130, no. 9, p. 103828, 2021

    2021

  22. [30]

    Safety helmet wearing detection based on yolov5 of attention mechanism,

    Z. Xu, Y . Zhang, J. Cheng, and G. Ge, “Safety helmet wearing detection based on yolov5 of attention mechanism,” in Journal of Physics: Conference Series, vol. 2213, no. 1. IOP Publishing, 2022, p. 012038

    2022

  23. [31]

    Multi-scale safety helmet detection based on sas-yolov3-tiny,

    R. Cheng, X. He, Z. Zheng, and Z. Wang, “Multi-scale safety helmet detection based on sas-yolov3-tiny,” Applied Sciences, vol. 11, no. 8, p. 3652, 2021

    2021

  24. [32]

    Work Safety Law of China (2021) , Standing Committee of the National People’s Congress, Jun. 2021

    2021

  25. [33]

    Condconv: Conditionally parameterized convolutions for efficient inference,

    B. Yang, G. Bender, J. Ngiam, and Q. V . Le, “Condconv: Conditionally parameterized convolutions for efficient inference,” 2019

    2019

  26. [34]

    Attention mechanisms in computer vision: A survey,

    M.-H. Guo, T.-X. Xu, J.-J. Liu, Z.-N. Liu, P.-T. Jiang, T.-J. Mu, S.- H. Zhang, R. R. Martin, M.-M. Cheng, and S.-M. Hu, “Attention mechanisms in computer vision: A survey,”Computational visual media, vol. 8, no. 3, pp. 331–368, 2022

    2022

  27. [35]

    Eca-net: Efficient channel attention for deep convolutional neural networks,

    Q. Wang, B. Wu, P. Zhu, P. Li, W. Zuo, and Q. Hu, “Eca-net: Efficient channel attention for deep convolutional neural networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 11 534–11 542

    2020

  28. [36]

    Non-local neural net- works,

    X. Wang, R. Girshick, A. Gupta, and K. He, “Non-local neural net- works,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 7794–7803

    2018

  29. [37]

    Coordinate attention for efficient mobile network design,

    Q. Hou, D. Zhou, and J. Feng, “Coordinate attention for efficient mobile network design,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 13 713–13 722

    2021

  30. [38]

    Strip pooling: Rethinking spatial pooling for scene parsing,

    Q. Hou, L. Zhang, M. M. Cheng, and J. Feng, “Strip pooling: Rethinking spatial pooling for scene parsing,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , 2020

    2020

  31. [39]

    Yolov6: A single-stage object detection framework for industrial applications,

    C. Li, L. Li, H. Jiang, K. Weng, Y . Geng, L. Li, Z. Ke, Q. Li, M. Cheng, W. Nie et al. , “Yolov6: A single-stage object detection framework for industrial applications,” arXiv preprint arXiv:2209.02976 , 2022

    2022 arXiv

  32. [40]

    Ghostnet: More features from cheap operations,

    K. Han, Y . Wang, Q. Tian, J. Guo, C. Xu, and C. Xu, “Ghostnet: More features from cheap operations,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 1580– 1589

    2020

  33. [41]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in 2009 IEEE conference on computer vision and pattern recognition . Ieee, 2009, pp. 248–255

    2009

  34. [42]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll ´ar, and C. L. Zitnick, “Microsoft coco: Common objects in context,” in Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V 13 . Springer,...

    2014

  35. [43]

    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 conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  36. [44]

    Visual transformers: Token- based image representation and processing for computer vision,

    B. Wu, C. Xu, X. Dai, A. Wan, P. Zhang, Z. Yan, M. Tomizuka, J. Gonzalez, K. Keutzer, and P. Vajda, “Visual transformers: Token- based image representation and processing for computer vision,” 2020

    2020

  37. [45]

    Yolov11: An overview of the key architectural enhancements,

    R. Khanam and M. Hussain, “Yolov11: An overview of the key architectural enhancements,” arXiv preprint arXiv:2410.17725 , 2024

    2024 arXiv

  38. [46]

    A generic shared attention mecha- nism for various backbone neural networks,

    Z. Huang, S. Liang, and M. Liang, “A generic shared attention mecha- nism for various backbone neural networks,” Neurocomputing, vol. 611, p. 128697, 2025

    2025

  39. [47]

    Mobilenetv2: Inverted residuals and linear bottlenecks,

    M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “Mobilenetv2: Inverted residuals and linear bottlenecks,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2018, pp. 4510–4520

    2018

  40. [48]

    Shufflenet v2: Practical guidelines for efficient cnn architecture design,

    N. Ma, X. Zhang, H.-T. Zheng, and J. Sun, “Shufflenet v2: Practical guidelines for efficient cnn architecture design,” in Proceedings of the European conference on computer vision (ECCV) , 2018, pp. 116–131

    2018

  41. [49]

    Rethinking bottleneck structure for efficient mobile network design,

    D. Zhou, Q. Hou, Y . Chen, J. Feng, and S. Yan, “Rethinking bottleneck structure for efficient mobile network design,” in Computer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part III 16 . Springer, 2020, pp. 680–697

    2020

  42. [50]

    Pp-lcnet: A lightweight cpu convolutional neural network,

    C. Cui, T. Gao, S. Wei, Y . Du, R. Guo, S. Dong, B. Lu, Y . Zhou, X. Lv, Q. Liu et al., “Pp-lcnet: A lightweight cpu convolutional neural network,” arXiv preprint arXiv:2109.15099 , 2021

    2021 arXiv

  43. [51]

    Aggregating global features into local vision transformer,

    K. Patel, A. M. Bur, F. Li, and G. Wang, “Aggregating global features into local vision transformer,” in 2022 26th International Conference on Pattern Recognition (ICPR). IEEE, 2022, pp. 1141–1147

    2022

  44. [52]

    Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting,

    Y . Zhang and J. Yan, “Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting,” in The eleventh international conference on learning representations , 2023

    2023

  45. [53]

    Vision transformer with deformable attention,

    Z. Xia, X. Pan, S. Song, L. E. Li, and G. Huang, “Vision transformer with deformable attention,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2022, pp. 4794–4803

    2022

  46. [54]

    Epsanet: An efficient pyramid split attention block on convolutional neural network,

    H. Zhang, K. Zu, J. Lu, Y . Zou, and D. Meng, “Epsanet: An efficient pyramid split attention block on convolutional neural network,” CoRR, 2021

    2021

  47. [55]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in European conference on computer vision . Springer, 2020, pp. 213– 229

    2020

  48. [56]

    Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,

    C. Y . Wang, A. Bochkovskiy, and H. Y . M. Liao, “Yolov7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors,” in arXiv, 2022

    2022

  49. [57]

    Fcos: Fully convolutional one- stage object detection,

    Z. Tian, C. Shen, H. Chen, and T. He, “Fcos: Fully convolutional one- stage object detection,” 2019

    2019

  50. [58]

    Pp-yoloe: An evolved version of yolo,

    S. Xu, X. Wang, W. Lv, Q. Chang, C. Cui, K. Deng, G. Wang, Q. Dang, S. Wei, and Y . Du, “Pp-yoloe: An evolved version of yolo,” 2022

    2022

  51. [59]

    Transformer for object detection: Review and benchmark,

    Y . Li, N. Miao, L. Ma, F. Shuang, and X. Huang, “Transformer for object detection: Review and benchmark,” Engineering Applications of Artificial Intelligence, vol. 126, p. 107021, 2023. IEEE TRANSACTIONS ON MOBILE COMPUTING 17

    2023

  52. [60]

    Edge-assisted real-time instance segmentation for resource-limited iot devices,

    Y . Xie, Y . Guo, Z. Mi, Y . Yang, and M. S. Obaidat, “Edge-assisted real-time instance segmentation for resource-limited iot devices,” IEEE Internet of Things Journal , vol. 10, no. 1, pp. 473–485, 2022

    2022

  53. [61]

    Iot edge computing-enabled collaborative tracking system for manufacturing re- sources in industrial park,

    Z. Zhao, P. Lin, L. Shen, M. Zhang, and G. Q. Huang, “Iot edge computing-enabled collaborative tracking system for manufacturing re- sources in industrial park,” Advanced Engineering Informatics , vol. 43, p. 101044, 2020

    2020

  54. [62]

    On the opportunities and risks of foundation models,

    R. Bommasani, D. A. Hudson, E. Adeli, R. Altman, S. Arora, S. von Arx, M. S. Bernstein, J. Bohg, A. Bosselut, E. Brunskill et al. , “On the opportunities and risks of foundation models,” arXiv preprint arXiv:2108.07258, 2021

    2021 arXiv

  55. [63]

    On the challenges and perspectives of foundation models for medical image analysis,

    S. Zhang and D. Metaxas, “On the challenges and perspectives of foundation models for medical image analysis,” Medical image analysis, vol. 91, p. 102996, 2024

    2024

Pith tools

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