Pith. sign in

REVIEW 5 major objections 6 minor 50 references

HyperDefect-YOLO: Enhance YOLO with HyperGraph Computation for Industrial Defect Detection

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 This paper claims that rebuilding a YOLO detector around hypergraph computation yields state-of-the-art accuracy on industrial defect detection while keeping inference fast enough for real-time use.

desk verdict A plausible engineering adaptation of HyperYOLO to defect detection with public code and honest ablations, but the SOTA claim rests on test-set-selected hyperparameters and uncontrolled baselines. read the letter →

arxiv 2412.03969 v1 pith:6U2BJDVZ submitted 2024-12-05 cs.CV

Zuo Zuo , Jiahao Dong , Yue Gao , Zongze Wu This is my paper · ORCID
classification cs.CV
keywords hypergraphcomputationindustrialdefectdetectionYOLOobjectmulti-scalefeaturefusionattentionmechanismPCBsteelsurfacedefectsreal-timeinspection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that embedding hypergraph computation into the YOLO pipeline yields a defect detector that outperforms prior general-purpose and task-specific detectors on industrial surface inspection. On the public HRIPCB printed-circuit-board benchmark it reports 98.1% precision and 98.2% mAP@0.5, and on the NEU-DET steel benchmark it reports 78.8% precision and 81.6% mAP@0.5, beating the hypergraph-based baseline, current YOLO versions, and dedicated defect detectors. The authors argue that hyperedges capture high-order feature interrelationships that plain convolution cannot express, which matters when defects are tiny, vary widely in scale, or sit inside a cluttered background. They also report practical operating costs: 5.4 ms inference, 9.3 million parameters, and 1.4 hours of training on NEU-DET. A sympathetic reader would take the paper as establishing that hypergraph modules can be integrated into a one-stage YOLO-style detector to improve both accuracy and speed for industrial defect detection.

What carries the argument

The carrying mechanism is hypergraph computation: image features are treated as vertices, similar features are grouped into hyperedges according to a distance threshold $\epsilon$, and hyperedge convolution (the matrix formulation in Eq. 2) propagates messages between vertices and their incident hyperedges, modeling high-order interrelationships that pairwise graphs cannot express. Around this core the paper builds four supporting modules. DAM generates defect attention maps to weaken background features; MGNet runs hypergraph and ordinary convolution branches in parallel; HGANet replaces hypergraph convolution with a distance-based attention map $\exp(-\mathrm{distance})$; CSF uses pixel unshuffle and attention-based fusion to avoid information loss during downsampling; and SAM gives the neck a mixture of receptive fields. Together these modules perform the central operation of the paper: concentrating on defect-relevant features while preserving tiny-defect information through the network.

What would settle it

Retrain HD-YOLO and the best baselines on NEU-DET and HRIPCB in one codebase with multiple random seeds; if the mAP@0.5 gap over the hypergraph baseline (0.9 on NEU-DET, 2.1 on HRIPCB) shrinks below seed variance, or if the rolled-in-scale collapse is not explained, the claim of uniformly state-of-the-art performance is falsified. A second check: fix $\epsilon=3$ for every dataset and evaluate on a held-out PCB test set, since the paper's own ablation shows that choice is optimal only for HRIPCB.

Watch

Extended reading notes

Core claim

The central claim is that rebuilding a YOLOv5-style detector around hypergraph computation produces state-of-the-art defect detection on three industrial datasets. The backbone replaces standard blocks with a Defect Aware Module (DAM), which applies sequential channel and spatial attention to suppress background, and a Mixed Graph Network (MGNet), which splits features into a hypergraph branch and a convolution branch to model high-order and low-order interrelationships in parallel. A HyperGraph Aggregation Network (HGANet) bridges backbone and neck, fusing multi-scale features with Distance-Based Attention, whose attention map is the negative exponential of the distance between visual features. In the neck, Cross-Scale Fusion (CSF) replaces concatenation-and-convolution fusion with pixel-unshuffle downsampling plus channel and spatial attention, and a Semantic Aware Module (SAM) mixes convolution kernels of sizes 1, 3, and 5 to enrich multi-scale semantics. With these modules, the paper reports the best precision and mAP@0.5 among all compared methods on HRIPCB, NEU-DET, and the self-built MINILED dataset, with the largest gains on tiny PCB defects and on the crazing and scratches classes of steel defects.

Load-bearing premise

The central claim depends on the assumption that the quoted baseline numbers were produced under comparable training settings and that single-run differences of 0.3 to 2.2 mAP points reflect real signal rather than run-to-run noise, especially because the hypergraph threshold and SAM kernel sizes were selected from test-set ablations.

Editorial extensions

If this is right

  • If the results hold, hypergraph-based feature interrelationships improve one-stage industrial defect detectors, extending the hypergraph-from-object-detection idea to manufacturing inspection.
  • The same module set works on PCB, steel strip, and mini-LED images, so a single architecture can be adapted to different production lines by retraining.
  • At 5.4 ms inference and 9.3M parameters on 640 by 640 inputs, the claimed accuracy would leave room for real-time automated optical inspection with additional post-processing.
  • The ablation progression (NEU-DET mAP@0.5 from 74.3 to 81.6) attributes most of the gain to DAM plus HGANet, indicating where future work on defect-specific attention should concentrate.
  • Dropping PANet in favor of FPN plus SAM implies that, for tiny defects, avoiding extra downsampling matters more than deeper feature fusion.

Reading between the lines

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

  • My inference: the hypergraph threshold $\epsilon$ and the SAM kernel set appear to be chosen per dataset from test-set ablations, so the fairest confirmation would fix them by cross-validation before comparing on new data.
  • My inference: the reported rolled-in-scale class collapse (58.3 versus 87.8 mAP@0.5 for the hypergraph baseline) shows the method is not uniformly better; identifying which appearance statistics cause hypergraph modeling to hurt would be a natural next step.
  • My inference: because HGANet relies on a distance-based attention map rather than learned hypergraph convolution, the same design could be extended to video or multi-view inspection by defining distance over spatio-temporal or cross-view features.
  • My inference: replacing the fixed distance threshold with a learned or per-image adaptive threshold would test whether the dataset-specific $\epsilon$ tuning can be eliminated.
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 HyperDefect-YOLO (HD-YOLO), a YOLOv5-based industrial defect detector that inserts hypergraph computation and attention modules: Defect Aware Module (DAM) and Mixed Graph Network (MGNet) in the backbone, HyperGraph Aggregation Network (HGANet) for multi-scale aggregation, and Cross-Scale Fusion (CSF) and Semantic Aware Module (SAM) in the neck. The authors evaluate on HRIPCB, NEU-DET, and a self-built MINILED dataset, reporting state-of-the-art Pre and mAP50 values with 5.4 ms inference and 9.3M parameters, and they provide a code link. The central claim is that HD-YOLO outperforms existing YOLO, transformer, and hypergraph detectors on these industrial datasets while remaining fast and lightweight.

Significance. If the empirical claims held, the paper would offer a useful speed-accuracy trade-off for industrial defect detection and demonstrate that hypergraph-based high-order feature modeling can be integrated into a lightweight YOLO detector. Strengths include evaluation on three datasets with different defect scales, internal ablations of each proposed module, efficiency measurements, and a public code link. However, the SOTA conclusion is not established by the current evidence, because the reported margins are smaller than the variation induced by test-set-selected hyperparameters, no variance or seed-repetition information is provided, and baseline numbers are not shown to come from identical training protocols.

major comments (5)
  1. [Section IV-B and IV-E, Table V] The hypergraph construction threshold epsilon is selected per dataset from the test set. Section IV-B fixes epsilon=3 for HRIPCB and epsilon=6 for NEU-DET and MINILED after Table V shows that these values give the best test mAP50. This makes the reported margins optimistic. On NEU-DET, changing epsilon from 6 to 3 lowers mAP50 from 81.6 to 79.3 (2.3 points), while the claimed advantage over HyperYOLO is only 0.9 points; on HRIPCB, threshold choice shifts mAP50 by 0.3-0.6 points, comparable to the 0.3-point margin over DRFA. The authors should select epsilon on a held-out validation split or fix it a priori, and then report the resulting performance.
  2. [Section IV-E, Table VI] The SAM kernel-size set {1,3,5} is likewise chosen from test-set ablations. On NEU-DET, the kernel set changes mAP50 from 78.2 to 81.6 and Pre from 72.9 to 80.7, so the selected configuration is the best on the test set by construction. Without a validation-based selection procedure, the comparison against fixed baseline models is not a fair test of the module's contribution.
  3. [Section IV-C, Tables I-III] Baseline numbers are quoted from previous publications or run externally, not reproduced under the identical training protocol used for HD-YOLO (input size, batch size, epochs, optimizer). For example, HRIPCB is evaluated at 1088x1088 with batch size 2 and 500 epochs, while no evidence is given that the baseline rows in Table I used the same settings. Because the reported advantages are 0.3-2.2 mAP points, differences in training protocol alone could explain the margins. A controlled reimplementation of all baselines under the same codebase and protocol is needed.
  4. [Section IV-C, Table II] On NEU-DET, HD-YOLO's per-class performance on 'Rs' (rolled-in scales) is 58.3 mAP50 versus 87.8 for HyperYOLO, a regression of 29.5 points. The paper acknowledges this drop but offers no analysis or remedy. This is inconsistent with the claim that HD-YOLO is uniformly superior and suggests the average advantage may be concentrated in particular classes. The authors should analyze the failure mode and report per-class results across multiple runs.
  5. [Section IV-D and IV-E] No error bars, seed repetitions, or statistical tests are reported for any of the tables. Since the decisive margins are as small as 0.3 mAP points, single-run results cannot be distinguished from training noise. The authors should report mean and standard deviation over at least three random seeds for the main comparison and ablations.
minor comments (6)
  1. [Section IV-C] The text 'depicted in Fig 7' should refer to Fig. 9, which shows the hard 'Rs' examples.
  2. [Table II] The HD-YOLO 'Sc' entries read '92. 9' and '80.4'; the spacing suggests a typo for 92.9.
  3. [Section III-C-2, Eq. (4)] The symbols Xle, Xconv1, Xconv2, Yconv2 and the channel-split operation are not defined with dimensions; please clarify.
  4. [Table IV] The checkmark rows are ambiguous because the active modules are not labeled; please replace with explicit configuration names such as 'baseline + DAM + HGANet'.
  5. [Section IV-D] The sentence stating MGNet 'suffers from performance deterioration on HRIPCB' is inconsistent with Table IV, where the all-module row outperforms the no-MGNet row on HRIPCB (98.2 vs 97.7 mAP50). Please clarify the intended comparison.
  6. [Figure 3] The CSF block shows multiple convolution kernels with k=3,s=2,p=1 and k=1,s=1,p=0; please annotate which branch each kernel belongs to.

Circularity Check

2 steps flagged · score 6.0 of 10

Reported SOTA margins on HRIPCB and NEU-DET are partially circular because the final hypergraph thresholds and SAM kernel sizes are selected on the same test sets whose metrics are then reported as predictions.

  1. fitted input called prediction [Section IV-B (Implementation Details) and Section IV-E (Ablation Studies, Table V)]
    "Hypergraph threshold is set to 3, 6, 6 for HRIPCB, NEU-DET and MINILED. ... Though P re are higher when ϵ is 6, we choose mAP0.5 as our main performance indicator. With regard to NEU-DET and MINILED, ϵ = 6 is the optimal hypergraph threshold observed from Table V."

    The final epsilon values are chosen by maximizing the same test-set metrics that are later reported as SOTA. On HRIPCB the metric itself is switched post hoc to justify epsilon=3; on NEU-DET, Table V shows epsilon=6 yields mAP0.5=81.6, the exact number reported as SOTA in Table II, while epsilon=5 yields 79.8. The claimed margins over HyperYOLO (0.9 on NEU-DET, 0.3 on HRIPCB) are smaller than the threshold-induced variation, so the reported 'prediction' is the selection outcome rather than an independent evaluation.

  2. fitted input called prediction [Section IV-E (Ablation Studies, Table VI)]
    "We set kernel sizes to {1, 3, 5 } in SAM. To validate the effectiveness of our chosen kernel sizes, we change the kernel sizes and conduct more experiments. The results are shown in Table VI. It can be found that the results are best on our chosen kernel sizes {1, 3, 5 }."

    The SAM receptive-field configuration is selected as the best row of a test-set sweep over kernel-size combinations. The final NEU-DET numbers (78.8 Pre, 81.6 mAP0.5) are exactly the values associated with {1,3,5}; other configurations swing mAP0.5 by 1.4-3.4 points (e.g., {1,5,7}: 78.2; {5,5,5}: 79.6), larger than several reported gains over baselines (e.g., +0.9 mAP0.5 vs HyperYOLO on NEU-DET). Thus the reported improvement is partly an artifact of selecting the kernel grid on the evaluation set itself.

full rationale

The architecture-level derivation is not circular: DAM, MGNet, HGANet, CSF and SAM are concrete modules with explicit equations, and the incremental ablations in Table IV are internally consistent. The self-citations to HGNN+ [24] and HyperYOLO [28] are used as component sources, not as proof of the present SOTA claim, so they do not by themselves make the derivation equivalent to its inputs. The central empirical claim, however, is partially circular in the evaluation-theoretic sense: the final hypergraph thresholds and SAM kernel sizes are chosen by scanning the same test sets whose SOTA numbers are then reported (Section IV-E, Tables V-VI), and one metric (mAP0.5) is adopted post hoc for HRIPCB. Consequently, the reported margins of 0.3-2.2 points are not independent predictions; they are the selected-best outcomes of the sweep. This is a fitted-input/called-prediction pattern rather than a self-citation chain. Additional concerns (no seed repetition or confidence intervals, baseline numbers quoted without a demonstrated common protocol, and the Rs-class collapse to 58.3 mAP50 vs 87.8 for HyperYOLO) are correctness risks, not circularity, and are not scored here.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central claim rests on empirical results rather than on formal axioms. The main tunable quantities are the hypergraph threshold and the SAM kernel set, both selected using test-set performance. No new physical or conceptual entities are introduced; all additions are network modules built from existing operations.

free parameters (2)
  • Hypergraph construction threshold epsilon = 3 (HRIPCB), 6 (NEU-DET, MINILED)
    Selected from test-set ablations (Table V); controls hypergraph edge density and materially changes results.
  • SAM convolution kernel size set = {1,3,5}
    Selected after test-set ablation (Table VI); other kernel sets drop mAP50 by 1.4 to 3.4 points on NEU-DET.
assumptions (3)
  • domain assumption Hypergraph convolution (HGNN+/HyperYOLO) models high-order feature interrelationships that improve defect detection.
    Central motivation; the only evidence is the ablation, and MGNet actually hurts HRIPCB, so the assumed benefit is not uniform.
  • domain assumption Pixel unshuffle is a lossless or near-lossless downsampling for tiny defects.
    Claimed in Section III-C4 and used as motivation for CSF; no information-loss measurement is provided.
  • domain assumption Defect features are out-of-distribution, making high-order relationships necessary.
    Stated in the Introduction; no distributional analysis is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HyperDefect-YOLO: Enhance YOLO with HyperGraph Computation for Industrial Defect Detection." pith.science (2026). https://pith.science/paper/6U2BJDVZ

@misc{pith2026241203969,
  author       = {Pith},
  title        = {Pith review of: HyperDefect-YOLO: Enhance YOLO with HyperGraph Computation for Industrial Defect Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6U2BJDVZ}},
  note         = {Machine review of arXiv:2412.03969}
}
read the original abstract

In the manufacturing industry, defect detection is an essential but challenging task aiming to detect defects generated in the process of production. Though traditional YOLO models presents a good performance in defect detection, they still have limitations in capturing high-order feature interrelationships, which hurdles defect detection in the complex scenarios and across the scales. To this end, we introduce hypergraph computation into YOLO framework, dubbed HyperDefect-YOLO (HD-YOLO), to improve representative ability and semantic exploitation. HD-YOLO consists of Defect Aware Module (DAM) and Mixed Graph Network (MGNet) in the backbone, which specialize for perception and extraction of defect features. To effectively aggregate multi-scale features, we propose HyperGraph Aggregation Network (HGANet) which combines hypergraph and attention mechanism to aggregate multi-scale features. Cross-Scale Fusion (CSF) is proposed to adaptively fuse and handle features instead of simple concatenation and convolution. Finally, we propose Semantic Aware Module (SAM) in the neck to enhance semantic exploitation for accurately localizing defects with different sizes in the disturbed background. HD-YOLO undergoes rigorous evaluation on public HRIPCB and NEU-DET datasets with significant improvements compared to state-of-the-art methods. We also evaluate HD-YOLO on self-built MINILED dataset collected in real industrial scenarios to demonstrate the effectiveness of the proposed method. The source codes are at https://github.com/Jay-zzcoder/HD-YOLO.

Figures

Figures reproduced from arXiv: 2412.03969 by the authors.

Figure 1
Figure 1. Examples and data distribution of different datasets. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The process of hypergraph construction. types of products as shown in [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Framework of the proposed HD-YOLO. B. Hypergraph learning methods Different from graph-based convolutional neural networks, the hypergraph neural network performs well in dealing with complex data correlation and multi-modal data representa￾tion. In hypergraph, the hyperedge connects more than two vertices which provides a more flexible way to represent and handle complex relationships. Hypergraph Neural Net￾works (… view at source ↗
Figures from the paper (9 more)
Figure 5
Figure 5. Figure 5: The structure of HyperGraph Aggregation Network (HGANet). [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: The scheme of pixel unshuffle. Multi-scale feature interaction and fusion can dramatically enrich semantic information and facilitate feature extraction and exploration. So before sending features in backbone to neck, we design HyperGraph Aggregation Network (HGANet) t…
Figure 7
Figure 7. Figure 7: Illustration of the proposed Semantic Aware Module (SAM). [PITH_FULL_IMAGE:figures/full_fig_p005_7.png]
Figure 9
Figure 9. Figure 9: Hard examples of defects in ”Rs” class of NEU-DET dataset. [PITH_FULL_IMAGE:figures/full_fig_p006_9.png]
Figure 6
Figure 6. Figure 6: After that, we concatenate two feature maps and use [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 10
Figure 10. Figure 10: Comparison of the SGD and Adam optimizers for HD-YOLO. [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 11
Figure 11. Figure 11: Results comparison of HD-YOLO and other detection methods on NEU-DET dataset. [PITH_FULL_IMAGE:figures/full_fig_p009_11.png]
Figure 12
Figure 12. Figure 12: Results comparison of HD-YOLO and other detection methods on HRIPCB dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_12.png]
Figure 13
Figure 13. Figure 13: Results comparison of HD-YOLO and other detection methods on MINILED dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_13.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 42 canonical work pages

  1. [1]

    Process manufacturing intelligence empowered by industrial JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 12 metaverse: A survey,

    W. Luo, K. Huang, X. Liang, H. Ren, N. Zhou, C. Zhang, C. Yang, and W. Gui, “Process manufacturing intelligence empowered by industrial JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 12 metaverse: A survey,” IEEE Trans. Cybern. , vol. 54, no. 11, pp. 6679– 6692, 2024

    work page 2020

  2. [2]

    Moninet with concurrent analytics of temporal and spatial information for fault detection in industrial processes,

    W. Yu, C. Zhao, and B. Huang, “Moninet with concurrent analytics of temporal and spatial information for fault detection in industrial processes,” IEEE Trans. Cybern. , vol. 52, no. 8, pp. 8340–8351, 2022

    work page 2022

  3. [3]

    Context-aware block net for small object detection,

    L. Cui, P. Lv, X. Jiang, Z. Gao, B. Zhou, L. Zhang, L. Shao, and M. Xu, “Context-aware block net for small object detection,” IEEE Trans. Cybern., vol. 52, no. 4, pp. 2300–2313, 2022

    work page 2022

  4. [4]

    Enhancing geometric factors in model learning and inference for object detection and instance segmentation,

    Z. Zheng, P. Wang, D. Ren, W. Liu, R. Ye, Q. Hu, and W. Zuo, “Enhancing geometric factors in model learning and inference for object detection and instance segmentation,” IEEE Trans. Cybern. , vol. 52, no. 8, pp. 8574–8586, 2022

    work page 2022

  5. [5]

    Taanet: A task-aware attention network for weak surface defect detection,

    L. Cui, S. Xie, E. Chen, X. Jiang, Z. Wang, X. Guo, and M. Xu, “Taanet: A task-aware attention network for weak surface defect detection,” IEEE Trans. Instrum. Meas. , vol. 73, pp. 1–12, 2024

    work page 2024

  6. [6]

    Clip-fsac: Boosting clip for few-shot anomaly classification with synthetic anomalies,

    Z. Zuo, Y . Wu, B. Li, J. Dong, Y . Zhou, L. Zhou, Y . Qu, and Z. Wu, “Clip-fsac: Boosting clip for few-shot anomaly classification with synthetic anomalies,” in Int. Joint Conf. Artif. Intell. , 2024, pp. 1834–1842

    work page 2024

  7. [7]

    A reconstruction-based feature adaptation for anomaly detection with self-supervised multi-scale aggregation,

    Z. Zuo, Z. Wu, B. Chen, and X. Zhong, “A reconstruction-based feature adaptation for anomaly detection with self-supervised multi-scale aggregation,” in IEEE Int. Conf. Acoust. Speech Signal Process. , 2024, pp. 5840–5844

    work page 2024

  8. [8]

    Ultralytics YOLOv8,

    G. Jocher, A. Chaurasia, and J. Qiu, “Ultralytics YOLOv8,” 2023. [Online]. Available: https://github.com/ultralytics/ultralytics

    2023

Show all 50 references
  1. [9]

    Faster r-cnn: Towards real-time object detection with region proposal networks,

    S. Ren, K. He, R. Girshick, and J. Sun, “Faster r-cnn: Towards real-time object detection with region proposal networks,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 39, no. 6, pp. 1137–1149, 2017

    2017

  2. [10]

    Detrs beat yolos on real-time object detection,

    Y . Zhao, W. Lv, S. Xu, J. Wei, G. Wang, Q. Dang, Y . Liu, and J. Chen, “Detrs beat yolos on real-time object detection,” in IEEE Conf. Comput. Vis. Pattern Recog., 2024, pp. 16 965–16 974

    2024

  3. [11]

    Yolo-hmc: An improved method for pcb surface defect detection,

    M. Yuan, Y . Zhou, X. Ren, H. Zhi, J. Zhang, and H. Chen, “Yolo-hmc: An improved method for pcb surface defect detection,” IEEE Trans. Instrum. Meas., vol. 73, pp. 1–11, 2024

    2024

  4. [12]

    Adin-detr: Adapting detection transformer for end-to-end real-time power line insulator defect detection,

    Y . Cheng and D. Liu, “Adin-detr: Adapting detection transformer for end-to-end real-time power line insulator defect detection,” IEEE Trans. Instrum. Meas., vol. 73, pp. 1–11, 2024

    2024

  5. [13]

    An optical lens defect detection method for micro vision based on wgso-yolo,

    Q. Lin, K. Takamasu, and M. Chen, “An optical lens defect detection method for micro vision based on wgso-yolo,” IEEE Trans. Instrum. Meas., vol. 73, pp. 1–8, 2024

    2024

  6. [14]

    Carafe: Content-aware reassembly of features,

    J. Wang, K. Chen, R. Xu, Z. Liu, C. C. Loy, and D. Lin, “Carafe: Content-aware reassembly of features,” in Int. Conf. Comput. Vis., 2019, pp. 3007–3016

    2019

  7. [15]

    An efficient anchor-free defect detector with dynamic receptive field and task alignment,

    F. Zuo, J. Liu, M. Fu, L. Wang, and Z. Zhao, “An efficient anchor-free defect detector with dynamic receptive field and task alignment,” IEEE Trans. Ind. Inform. , vol. 20, pp. 8536–8547, 2024

    2024

  8. [16]

    Mci-gla plug-in suitable for yolo series models for transmission line insulator defect detection,

    Y . Wang, X. Song, L. Feng, Y . Zhai, Z. Zhao, S. Zhang, and Q. Wang, “Mci-gla plug-in suitable for yolo series models for transmission line insulator defect detection,” IEEE Trans. Instrum. Meas. , vol. 73, pp. 1–12, 2024

    2024

  9. [17]

    Hripcb: a challenging dataset for pcb defects detection and classification,

    H. Weibo, W. Peng, Z. Manhua, and H. Liu, “Hripcb: a challenging dataset for pcb defects detection and classification,” J. Eng., vol. 2020, no. 13, p. 303–309, 2020

    2020

  10. [18]

    An end-to-end steel surface defect detection approach via fusing multiple hierarchical features,

    Y . He, K. Song, Q. Meng, and Y . Yan, “An end-to-end steel surface defect detection approach via fusing multiple hierarchical features,” IEEE Trans. Instrum. Meas. , vol. 69, no. 4, pp. 1493–1504, 2020

    2020

  11. [19]

    Multilevel fine- grained features-based general framework for object detection,

    F. Zuo, J. Liu, Z. Chen, H. Zhang, M. Fu, and L. Wang, “Multilevel fine- grained features-based general framework for object detection,” IEEE Trans. Cybern., vol. 54, no. 11, pp. 6921–6933, 2024

    2024

  12. [20]

    Joining spatial deformable convolution and a dense feature pyramid for surface defect detection,

    Z. Huang, C. Zhang, L. Ge, Z. Chen, K. Lu, and C. Wu, “Joining spatial deformable convolution and a dense feature pyramid for surface defect detection,” IEEE Trans. Instrum. Meas. , vol. 73, pp. 1–14, 2024

    2024

  13. [21]

    Pcb-yolo: An improved detection algorithm of pcb surface defects based on yolov5,

    J. Tang, S. Liu, D. Zhao, L. Tang, W. Zou, and B. Zheng, “Pcb-yolo: An improved detection algorithm of pcb surface defects based on yolov5,” Sustainability, vol. 15, no. 7, 2023

    2023

  14. [22]

    Canet: Contextual information and spatial attention based network for detecting small defects in manufacturing industry,

    X. Hou, M. Liu, S. Zhang, P. Wei, and B. Chen, “Canet: Contextual information and spatial attention based network for detecting small defects in manufacturing industry,” Pattern Recognition , vol. 140, p. 109558, 2023

    2023

  15. [23]

    Hypergraph neural networks,

    Y . Feng, H. You, Z. Zhang, R. Ji, and Y . Gao, “Hypergraph neural networks,” in AAAI, 2019

    2019

  16. [24]

    HGNN+: General hypergraph neural networks,

    Y . Gao, Y . Feng, S. Ji, and R. Ji, “HGNN+: General hypergraph neural networks,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 45, pp. 3181– 3199, 2023

    2023

  17. [25]

    Lbsn2vec++: Het- erogeneous hypergraph embedding for location-based social networks,

    D. Yang, B. Qu, J. Yang, and P. Cudr ´e-Mauroux, “Lbsn2vec++: Het- erogeneous hypergraph embedding for location-based social networks,” IEEE Trans. Knowl. Data Eng. , vol. 34, no. 4, pp. 1843–1855, 2022

    2022

  18. [26]

    Hypergraph factorization for multi-tissue gene expression imputation,

    V . Ramon, K. J. Chaitanya, G. Dobrik, L. Phillip, D. Bianca, R. G. Eric, and L. Pietro, “Hypergraph factorization for multi-tissue gene expression imputation,” Nature Machine Intelligence , vol. 5, no. 7, p. 739–753, 2023

    2023

  19. [27]

    Multi-hypergraph learning- based brain functional connectivity analysis in fmri data,

    L. Xiao, J. Wang, P. H. Kassani, Y . Zhang, Y . Bai, J. M. Stephen, T. W. Wilson, V . D. Calhoun, and Y .-P. Wang, “Multi-hypergraph learning- based brain functional connectivity analysis in fmri data,” IEEE Trans. Med. Imag., vol. 39, no. 5, pp. 1746–1758, 2020

    2020

  20. [28]

    Hyper-yolo: When visual object detection meets hypergraph computation,

    Y . Feng, J. Huang, S. Du, S. Ying, J.-H. Yong, Y . Li, G. Ding, R. Ji, and Y . Gao, “Hyper-yolo: When visual object detection meets hypergraph computation,” ArXiv Preprint arXiv:2408.04804 , 2024

    2024 arXiv

  21. [29]

    You only look once: Unified, real-time object detection,

    J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” in IEEE Conf. Comput. Vis. Pattern Recog., 2016, pp. 779–788

    2016

  22. [30]

    Yolo9000: Better, faster, stronger,

    J. Redmon and A. Farhadi, “Yolo9000: Better, faster, stronger,” in IEEE Conf. Comput. Vis. Pattern Recog. , 2017, pp. 6517–6525

    2017

  23. [31]

    Yolox: Exceeding yolo series in 2021,

    Z. Ge, S. Liu, F. Wang, Z. Li, and J. Sun, “Yolox: Exceeding yolo series in 2021,” ArXiv Preprint arXiv:2107.08430 , 2021

    2021 arXiv

  24. [32]

    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, Y . Li, B. Zhang, Y . Liang, L. Zhou, X. Xu, X. Chu, X. Wei, and X. Wei, “Yolov6: A single-stage object detection framework for industrial applications,” ArXiv Preprint arXiv:2209.02976 , 2022

    2022 arXiv

  25. [33]

    Repvgg: Making vgg-style convnets great again,

    X. Ding, X. Zhang, N. Ma, J. Han, G. Ding, and J. Sun, “Repvgg: Making vgg-style convnets great again,” in IEEE Conf. Comput. Vis. Pattern Recog., 2021, pp. 13 728–13 737

    2021

  26. [34]

    Tood: Task- aligned one-stage object detection,

    C. Feng, Y . Zhong, Y . Gao, M. R. Scott, and W. Huang, “Tood: Task- aligned one-stage object detection,” in 2021 IEEE/CVF International Conference on Computer Vision (ICCV) , 2021, pp. 3490–3499

    2021

  27. [35]

    Yolov9: Learning what you want to learn using programmable gradient information,

    C.-Y . Wang, I.-H. Yeh, and H.-Y . M. Liao, “Yolov9: Learning what you want to learn using programmable gradient information,” ArXiv Preprint arXiv:2402.13616, 2024

    2024 arXiv

  28. [36]

    Yolov10: Real-time end-to-end object detection,

    A. Wang, H. Chen, L. Liu, K. Chen, Z. Lin, J. Han, and G. Ding, “Yolov10: Real-time end-to-end object detection,” ArXiv Preprint arXiv:2405.14458, 2024

    2024 arXiv

  29. [37]

    Ultralytics YOLOv11,

    “Ultralytics YOLOv11,” 2024. [Online]. Available: https://github.com/ ultralytics/ultralytics

    2024

  30. [38]

    Real-time single image and video super- resolution using an efficient sub-pixel convolutional neural network,

    W. Shi, J. Caballero, F. Husz ´ar, J. Totz, A. P. Aitken, R. Bishop, D. Rueckert, and Z. Wang, “Real-time single image and video super- resolution using an efficient sub-pixel convolutional neural network,” in IEEE Conf. Comput. Vis. Pattern Recog. , 2016, pp. 1874–1883

    2016

  31. [39]

    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,” ArXiv Preprint arXiv:2207.02696 , 2022

    2022 arXiv

  32. [40]

    Detectors: Detecting objects with recursive feature pyramid and switchable atrous convolution,

    S. Qiao, L.-C. Chen, and A. Yuille, “Detectors: Detecting objects with recursive feature pyramid and switchable atrous convolution,” in IEEE Conf. Comput. Vis. Pattern Recog. , 2021, pp. 10 208–10 219

    2021

  33. [41]

    Automatic detection and counting system for pavement cracks based on pcgan and yolo-mf,

    D. Ma, H. Fang, N. Wang, C. Zhang, J. Dong, and H. Hu, “Automatic detection and counting system for pavement cracks based on pcgan and yolo-mf,” IEEE Trans. Intell. Transp. Syst. , vol. 23, no. 11, pp. 22 166– 22 178, 2022

    2022

  34. [42]

    Es-net: Efficient scale- aware network for tiny defect detection,

    X. Yu, W. Lyu, D. Zhou, C. Wang, and W. Xu, “Es-net: Efficient scale- aware network for tiny defect detection,” IEEE Trans. Instrum. Meas. , vol. 71, pp. 1–14, 2022

    2022

  35. [43]

    Deformable yolox: Detection and rust warning method of transmission line connection fittings based on image processing technology,

    Z. Song, X. Huang, C. Ji, and Y . Zhang, “Deformable yolox: Detection and rust warning method of transmission line connection fittings based on image processing technology,” IEEE Trans. Instrum. Meas. , vol. 72, pp. 1–21, 2023

    2023

  36. [44]

    An anchor-free defect detector for complex background based on pixelwise adaptive multiscale feature fusion,

    H. Lu, M. Fang, Y . Qiu, and W. Xu, “An anchor-free defect detector for complex background based on pixelwise adaptive multiscale feature fusion,” IEEE Trans. Instrum. Meas. , vol. 72, pp. 1–12, 2023

    2023

  37. [45]

    Attention network for rail surface defect detection via consistency of intersection-over-union(iou)- guided center-point estimation,

    X. Ni, Z. Ma, J. Liu, B. Shi, and H. Liu, “Attention network for rail surface defect detection via consistency of intersection-over-union(iou)- guided center-point estimation,” IEEE Trans. Ind. Inform., vol. 18, no. 3, pp. 1694–1705, 2022

    2022

  38. [46]

    Visual fault detection of multiscale key components in freight trains,

    Y . Zhang, Y . Zhou, H. Pan, B. Wu, and G. Sun, “Visual fault detection of multiscale key components in freight trains,” IEEE Trans. Ind. Inform. , vol. 19, no. 8, pp. 9082–9090, 2023

    2023

  39. [47]

    Cascade r-cnn: High quality object detection and instance segmentation,

    Z. Cai and N. Vasconcelos, “Cascade r-cnn: High quality object detection and instance segmentation,” IEEE Trans. Pattern Anal. Mach. Intell. , vol. 43, no. 5, pp. 1483–1498, 2021

    2021

  40. [48]

    Libra r-cnn: Towards balanced learning for object detection,

    J. Pang, K. Chen, J. Shi, H. Feng, W. Ouyang, and D. Lin, “Libra r-cnn: Towards balanced learning for object detection,” in IEEE Conf. Comput. Vis. Pattern Recog., 2019, pp. 821–830

    2019

  41. [49]

    Focal loss for dense object detection,

    T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” in Int. Conf. Comput. Vis., 2017, pp. 2999–3007

    2017

  42. [50]

    Fcos: Fully convolutional one- stage object detection,

    Z. Tian, C. Shen, H. Chen, and T. He, “Fcos: Fully convolutional one- stage object detection,” in Int. Conf. Comput. Vis., 2019, pp. 9626–9635

    2019

Pith tools

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