Pith. sign in

REVIEW 3 major objections 6 minor 50 references

J-DDL: Surface Damage Detection and Localization System for Fighter Aircraft

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read A rail-mounted scan-and-laser system can detect surface damage on fighter aircraft in 2D and map each defect to its 3D location on the aircraft body.

desk verdict Useful dataset and platform, but the 3D localization claim is unvalidated and Algorithm 1 has a non-invertible matrix problem. read the letter →

arxiv 2506.10505 v1 pith:AZMMYZ6Y submitted 2025-06-12 cs.CV

classification cs.CV
keywords aircraftsurfacedamagedetection3DlocalizationpointcloudmappingYOLOmulti-scaleattentiondatasetinspectionInner-CIoUloss
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 proposes J-DDL, a rail-based inspection system for fighter aircraft that combines cameras and laser scanners to capture the full surface as 2D images and 3D point clouds. The central claim is that damage detected in the images can be mapped accurately onto the point cloud, giving 3D locations of defects. At the heart of the detection stage is AIR-YOLO, a modified YOLO detector that reports 65.2% mAP at 110 FPS on the authors' new AIRSD dataset of 8,091 images covering 11 damage types. The paper further reports an operational comparison in which the system found 87 defects in 28 minutes, versus 31 defects in 3.1 hours for a manual inspector. If these figures hold, automated hangar inspections could replace or prioritize lengthy human visual checks.

What carries the argument

The load-bearing object is the projection-and-back-projection pipeline defined in Algorithm 1: every laser point is projected into the image plane using the camera's intrinsics $K$ and extrinsics $[R|T]$, points that fall inside a detected damage bounding box are kept, and those kept points are back-projected to 3D using their stored depth. This single mechanism converts a 2D detection (a bounding box) into an explicit 3D region without any learned geometry. The detection side is carried by three named components: FasterNet blocks that cut backbone parameters by about 30%, EMA attention modules that add 0.9% mAP, and the Inner-CIoU loss that outperforms CIoU, GIoU, and DIoU on the same benchmark.

What would settle it

Take a physical aircraft with a defect of precisely known 3D coordinates (established with a coordinate-measuring machine or survey total station), run the J-DDL capture and the Algorithm-1 mapping, and measure the distance between the reported damaged point-cloud cluster and the true coordinate; if the distance is large, or if the point cloud has no return at the defect location, the claimed accurate 3D localization fails. A second check is to test the same dataset with camera intrinsics perturbed by a few pixels to see how quickly the back-projected damage points move off the true defect.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a non-contact rail platform can, in a single pass, detect and three-dimensionally localize visible surface damage on a fighter aircraft by fusing 2D detections with laser-scanned point clouds. For the 2D stage, the authors introduce AIR-YOLO, which lightens the YOLOv8 backbone with FasterNet-style partial convolutions, adds Efficient Multi-scale Attention (EMA) modules to the neck, and replaces CIoU with an Inner-CIoU loss; ablation shows the full combination reaches 65.2% mAP on the AIRSD test set. For the 3D stage, they project each detected damage bounding box through the camera projection matrix $P=K[R|T]$ to select the corresponding point-cloud points, producing a localized 3D damage region.

Load-bearing premise

The 3D localization step assumes the camera intrinsics and extrinsics give a correct projection matrix and that the laser point cloud is dense and accurate at every damage location; the paper provides no calibration procedure, no point-cloud density check, and no localization-error measurement, so an imperfect calibration or a depth hole at the defect would invalidate the localization even when the 2D detection is correct.

Editorial extensions

If this is right

  • A single hangar pass can localize 87 defects in 28 minutes, versus 31 defects in 3.1 hours by a manual inspector, per the paper's Table V.
  • The 2D detector runs at 110 FPS, so image analysis does not bottleneck the scanning rate.
  • AIRSD, the first public fighter-aircraft damage dataset with 8,091 images and 11 damage types, gives other teams a common benchmark.
  • Per-category results suggest strong performance for missing fastener (mAP 82.1%), lightning strike (mAP 99.5%), and rivet damage (mAP 71.6%), while scratch (39.5%) and paint peeling (56.0%) remain weak spots.
  • Because the system is non-contact and rail-mounted, it avoids the collision risk of UAV-based inspection of aircraft surfaces.

Reading between the lines

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

  • The paper reports no measurement of 3D localization error, so the 'precise localization' claim is not yet quantified; a reader could test it by placing a marker at a known coordinate and comparing the system's output to a survey-grade measurement.
  • The same projection-mapping scheme should transfer to other large curved structures, such as transport aircraft, ship hulls, or wind-turbine blades, provided the camera calibration and point-cloud density hold.
  • The synthetic-plus-real training with a real-only test set invites a follow-up that measures how much synthetic data actually harms or helps each damage category; the current paper does not report that split.
  • The three architectural changes are each evaluated only inside the combined AIR-YOLO, so the individual contributions might not transfer identically to other detectors, but the Inner-CIoU loss's improvement over CIoU is consistent with its reported ablation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper presents J-DDL, a rail-based inspection system for fighter aircraft that combines 2D cameras and 3D laser scanners to detect and localize surface damage. The detection component, AIR-YOLO, modifies YOLOv8 with Fasternet backbone blocks, EMA attention modules in the neck, and an Inner-CIoU loss function. The authors also introduce AIRSD, a new dataset of 8,091 images covering 11 damage categories with both synthetic and real images. Detection experiments on AIRSD report a mAP of 65.2% at 110 FPS, and a qualitative 3D localization example is shown. The central claim is that the system achieves precise damage detection and 3D localization across the entire aircraft surface.

Significance. The system-level goal is practically important: an automated, non-contact inspection of fighter aircraft surfaces could reduce inspection time from hours to tens of minutes. The AIRSD dataset is a useful community resource, especially because the test set is reportedly real-world images. The detection experiments are reproduced with several baselines and an ablation table, and the reported gains are modest but internally plausible. However, the 3D localization half of the central claim is not quantitatively validated, and the manual-inspection comparison does not support the accuracy claim. As presented, the paper supports a moderate incremental contribution in 2D damage detection but not the full system claim.

major comments (3)
  1. [Section III.E and Algorithm 1] The damage localization procedure is not executable as written because it uses the inverse of the projection matrix P, which is a 3x4 matrix and therefore has no inverse. The step 'P^{-1}' in Algorithm 1 must be replaced by a well-defined back-projection that accounts for the depth Z (e.g., an explicit least-squares or pseudo-inverse formulation). More importantly, the paper provides no camera calibration procedure, no point-cloud density or accuracy guarantee, and no localization error metric. The only localization evidence is one qualitative point-cloud example in Fig. 7. Since the abstract and conclusion explicitly claim 'accurate 3D localization,' this is a load-bearing gap; the authors should add a calibration protocol, quantify reprojection error, and evaluate localization accuracy against ground-truth 3D positions or physical measurements.
  2. [Section V.B, Table V] The comparison with manual inspection reports 87 detected defects by the proposed method versus 31 by an experienced inspector, and times of 28 minutes versus 3.1 hours. Without a ground-truth reference, the higher count cannot be interpreted as higher accuracy; it may include many false positives, and the single manual inspection is not a statistically reliable baseline. The sentence 'our framework also improves accuracy' is therefore unsupported. The authors should report precision and recall against a verified defect list, or at least provide a blind cross-check by a second inspector.
  3. [Section IV.D and Fig. 6] The claim that AIR-YOLO achieves 65.2% mAP at 110 FPS on the AIRSD test set is presented without reporting the test-set size, the confidence threshold used for FPS, or the hardware configuration for inference. More importantly, no per-class or per-image uncertainty measures are provided. For a safety-critical application, reporting mAP alone is not sufficient; the authors should add confidence-interval estimates or error analysis, particularly for the low-performing classes (scratch mAP 39.5%, paint peeling mAP 56.0%) already identified in Table IV.
minor comments (6)
  1. [Section III.D and Eq. (8)] The ratio parameter in Inner-CIoU, described as typically in [0.5, 1.5], is a free hyperparameter but its value is not reported anywhere in the experiments; please state the value used and justify the choice.
  2. [Section II.A and Section III.B] There is a typo: 'Fasterent [40]' should read 'FasterNet [40]'.
  3. [Section IV.A] The dataset link is a Google Drive folder; for a 'publicly available dataset' claim, please provide a persistent DOI or a more stable repository, and state the exact number of test images.
  4. [Section IV.A] The statement that the test set is exclusively real-world images is not accompanied by the train/validation/test split; please report the number of images per split and confirm that no synthetic images appear in the test set.
  5. [Section IV.C, Table II] The ablation table would be clearer if each row explicitly indicated which component was added to the baseline; as printed, the checkmark layout makes it easy to misread the single-component rows as full configurations.
  6. [Section III.D] Equation (8) uses IoU and IoU_inner without defining IoU in the text; please define the standard IoU term or cite the original Inner-IoU formulation precisely.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: detection components are externally cited and empirically evaluated on a new dataset; the localization flaw is a correctness risk, not a circularity.

full rationale

I walked the claimed derivation chain. The 2D detection network is YOLOv8 augmented with external modules: FasterNet blocks ([40]), EMA attention ([41]), and Inner-IoU ([42]); each is explicitly cited, and each is ablated or compared against baselines retrained under the same protocol on the new AIRSD dataset, whose test split is real-world. The reported mAP is therefore an empirical result on held-out data, not a fitted parameter renamed as a prediction. The 3D localization procedure uses the standard projection P=K[R|T] and point-cloud back-projection; Algorithm 1's notation P^{-1} for a non-square 3x4 matrix is mathematically invalid and the localization error is not measured, but these are correctness and validation gaps, not circular reductions. Self-citations ([4]-[6], [29]) appear only as related work and carry no load-bearing argument. The unreported Inner-CIoU Ratio hyperparameter is a standard tuning choice, and the paper's 'novel' label for Inner-CIoU is an attribution/novelty concern rather than circularity because the loss is directly imported from an external source and its benefit is empirically tested. No step equates the output to its input by construction or via a self-citation chain.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central detection claim depends on known modules and standard training, with one unreported hyperparameter (Ratio). The localization claim rests on unverified calibration, full surface coverage, annotation quality, and synthetic-to-real transfer. No new entities are introduced.

free parameters (1)
  • Inner-CIoU Ratio = not reported (claimed range [0.5, 1.5])
    Scaling factor for auxiliary bounding boxes in Eq. (1)-(8). The paper says the value is typically in [0.5, 1.5] but does not report the value used in experiments or a sweep; the loss function and resulting mAP depend on it.
assumptions (4)
  • domain assumption Camera intrinsics and extrinsics are accurately known and stable, giving a correct projection matrix P=K[R|T] that aligns 3D points to 2D pixels
    Algorithm 1 projects point clouds through P and back-projects by inverse P. No calibration procedure, accuracy, or stability analysis is reported in Section III.E, so this assumption is load-bearing for the 3D localization claim.
  • domain assumption The rail-mounted platform covers the entire aircraft surface at sufficient resolution for all defects that matter
    Sections I and V assert complete coverage, but Section IV.F admits limited resolution can miss small defects, so the coverage assumption has known exceptions that weaken the 'entire surface' claim.
  • domain assumption The AIRSD test set is real-world, correctly annotated, and representative of operational damage
    Section IV.A states test is exclusively real images annotated with X-AnyLabeling, but no annotation agreement or dataset statistics are given, and the class distribution is heavily imbalanced.
  • domain assumption Synthetic training images are close enough to real surfaces for the model to generalize
    Section IV.A trains on synthetic overlays while testing on real images; Section IV.F explicitly states synthetic images may not replicate real-world lighting and texture, making this a stated-but-unvalidated assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of J-DDL: Surface Damage Detection and Localization System for Fighter Aircraft." pith.science (2026). https://pith.science/paper/AZMMYZ6Y

@misc{pith2026250610505,
  author       = {Pith},
  title        = {Pith review of: J-DDL: Surface Damage Detection and Localization System for Fighter Aircraft},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AZMMYZ6Y}},
  note         = {Machine review of arXiv:2506.10505}
}
read the original abstract

Ensuring the safety and extended operational life of fighter aircraft necessitates frequent and exhaustive inspections. While surface defect detection is feasible for human inspectors, manual methods face critical limitations in scalability, efficiency, and consistency due to the vast surface area, structural complexity, and operational demands of aircraft maintenance. We propose a smart surface damage detection and localization system for fighter aircraft, termed J-DDL. J-DDL integrates 2D images and 3D point clouds of the entire aircraft surface, captured using a combined system of laser scanners and cameras, to achieve precise damage detection and localization. Central to our system is a novel damage detection network built on the YOLO architecture, specifically optimized for identifying surface defects in 2D aircraft images. Key innovations include lightweight Fasternet blocks for efficient feature extraction, an optimized neck architecture incorporating Efficient Multiscale Attention (EMA) modules for superior feature aggregation, and the introduction of a novel loss function, Inner-CIOU, to enhance detection accuracy. After detecting damage in 2D images, the system maps the identified anomalies onto corresponding 3D point clouds, enabling accurate 3D localization of defects across the aircraft surface. Our J-DDL not only streamlines the inspection process but also ensures more comprehensive and detailed coverage of large and complex aircraft exteriors. To facilitate further advancements in this domain, we have developed the first publicly available dataset specifically focused on aircraft damage. Experimental evaluations validate the effectiveness of our framework, underscoring its potential to significantly advance automated aircraft inspection technologies.

Figures

Figures reproduced from arXiv: 2506.10505 by the authors.

Figure 1
Figure 1. Overview of our proposed framework. The designed platform captures 2D images and 3D point clouds of the airplane. Next, the AIR-YOLO trained [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Architecture of the proposed AIR-YOLO network. Key enhancements include a lightweight backbone incorporating Fasternet blocks comprising [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Structure of EMA attention module. C. Feature Aggregation Given that the background information in airplane surface images is inherently less complex than that of natural images, we introduce Efficient Multi-scale Attention (EMA) modules into the neck layer of our network. These modules are strate￾gically positioned downstream of the C2f module to optimize the aggregation of multi-scale features. By leveraging the E… view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Damage detection results for airplane surface. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison between AIR-YOLO and other SOTA models. [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: 3D localization of damage on the airplane surface. (a) 3D point cloud [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Our hardware platform. Cameras and scanners are mounted on the mechanical supports, enabling the acquisition of both 2D images and 3D point [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 36 canonical work pages

  1. [1]

    Aircraft visual inspection: A systematic literature review,

    Y . D. Yasuda, F. A. Cappabianco, L. E. G. Martins, and J. A. Gripp, “Aircraft visual inspection: A systematic literature review,”Computers in Industry, vol. 141, p. 103695, 2022

  2. [2]

    Preliminary design of an unmanned aircraft system for aircraft general visual inspection,

    U. Papa and S. Ponte, “Preliminary design of an unmanned aircraft system for aircraft general visual inspection,”Electronics, vol. 7, no. 12, p. 435, 2018

  3. [3]

    Human factors in maintenance: a review,

    M. Sheikhalishahi, L. Pintelon, and A. Azadeh, “Human factors in maintenance: a review,”Journal of Quality in Maintenance Engineering, vol. 22, no. 3, pp. 218–237, 2016

  4. [4]

    Tensor voting guided mesh denoising,

    M. Wei, L. Liang, W.-M. Pang, J. Wang, W. Li, and H. Wu, “Tensor voting guided mesh denoising,”IEEE Transactions on Automation Science and Engineering, vol. 14, no. 2, pp. 931–945, 2017. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 13

  5. [5]

    Multiscale feature line extraction from raw point clouds based on local surface variation and anisotropic contraction,

    H. Chen, Y . Huang, Q. Xie, Y . Liu, Y . Zhang, M. Wei, and J. Wang, “Multiscale feature line extraction from raw point clouds based on local surface variation and anisotropic contraction,”IEEE Transactions on Automation Science and Engineering, vol. 19, no. 2, pp. 1003–1016, 2022

  6. [6]

    City3d: Large-scale build- ing reconstruction from airborne lidar point clouds,

    J. Huang, J. Stoter, R. Peters, and L. Nan, “City3d: Large-scale build- ing reconstruction from airborne lidar point clouds,”Remote Sensing, vol. 14, no. 9, 2022

  7. [7]

    A uav-based aircraft surface defect inspection system via external constraints and deep learning,

    Y . Liu, J. Dong, Y . Li, X. Gong, and J. Wang, “A uav-based aircraft surface defect inspection system via external constraints and deep learning,”IEEE Transactions on Instrumentation and Measurement, vol. 71, pp. 1–15, 2022

  8. [8]

    Aircraft skin damage detection and assessment from uav images using glcm and cloud model,

    L. Shao, J. He, X. Lu, B. Hei, J. Qu, and W. Liu, “Aircraft skin damage detection and assessment from uav images using glcm and cloud model,” IEEE Transactions on Intelligent Transportation Systems, 2023

Show all 50 references
  1. [9]

    A semi-supervised aircraft fuselage defect detection network with dynamic attention and class-aware adaptive pseudo-label assignment,

    X. Zhang, J. Zhang, J. Chen, R. Guo, and J. Wu, “A semi-supervised aircraft fuselage defect detection network with dynamic attention and class-aware adaptive pseudo-label assignment,”IEEE Transactions on Artificial Intelligence, 2024

  2. [10]

    Ultralytics YOLO,

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

  3. [11]

    Applications of nde reliability to systems,

    W. Rummel, G. Hardy, and T. Cooper, “Applications of nde reliability to systems,”ASM Handbook., vol. 17, pp. 674–688, 1989

  4. [12]

    Reliability assessment at airline in- spection facilities. volume iii: Results of an eddy current inspection reliability experiment,

    F. Spencer and D. Schurman, “Reliability assessment at airline in- spection facilities. volume iii: Results of an eddy current inspection reliability experiment,”Department of Transportation, Federal Aviation Administration, Washington, DC, 1995

  5. [13]

    Murgatroyd, G

    R. Murgatroyd, G. Worrall, and C. Waites,A study of the human factors influencing the reliability of aircraft inspection. Civil Aviation Authority, 1995

  6. [14]

    A review of human error in aviation maintenance and inspection,

    K. A. Latorella and P. V . Prabhu, “A review of human error in aviation maintenance and inspection,”Human Error in Aviation, pp. 521–549, 2017

  7. [15]

    A mobile climbing robot for high precision manufacture and inspection of aerostructures,

    T. S. White, R. Alexander, G. Callow, A. Cooke, S. Harris, and J. Sargent, “A mobile climbing robot for high precision manufacture and inspection of aerostructures,”The International Journal of Robotics Research, vol. 24, no. 7, pp. 589–598, 2005

  8. [16]

    Design of a climbing robot for inspecting aircraft wings and fuselage,

    J. Shang, T. Sattar, S. Chen, and B. Bridge, “Design of a climbing robot for inspecting aircraft wings and fuselage,”Industrial Robot: An International Journal, vol. 34, no. 6, pp. 495–502, 2007

  9. [17]

    Collaborative robot for visually inspecting an aircraft,

    N. Colin and F. Guibert, “Collaborative robot for visually inspecting an aircraft,” Apr. 24 2018, uS Patent 9,952,593

  10. [18]

    A navigational framework combining visual servoing and spiral obstacle avoidance techniques,

    M. Futterlieb, V . Cadenat, and T. Sentenac, “A navigational framework combining visual servoing and spiral obstacle avoidance techniques,” in2014 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO), vol. 2. IEEE, 2014, pp. 57–64

  11. [19]

    Moving obsta- cles detection and camera pointing for mobile robot applications,

    M. Lakrouf, S. Larnier, M. Devy, and N. Achour, “Moving obsta- cles detection and camera pointing for mobile robot applications,” in Proceedings of the 3rd International Conference on Mechatronics and Robotics Engineering, 2017, pp. 57–62

  12. [20]

    Rich feature hierarchies for accurate object detection and semantic segmentation,

    R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2014, pp. 580–587

  13. [21]

    Fast r-cnn,

    R. Girshick, “Fast r-cnn,” inProceedings of the IEEE international conference on computer vision, 2015, pp. 1440–1448

  14. [22]

    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,”Advances in neural information processing systems, vol. 28, 2015

  15. [23]

    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,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779– 788

  16. [24]

    Yolov4: Op- timal speed and accuracy of object detection,

    A. Bochkovskiy, C.-Y . Wang, and H.-Y . M. Liao, “Yolov4: Op- timal speed and accuracy of object detection,”arXiv preprint arXiv:2004.10934, 2020

  17. [25]

    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. Nieet al., “Yolov6: A single-stage object detection framework for industrial applications,”arXiv preprint arXiv:2209.02976, 2022

  18. [26]

    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,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 7464–7475

  19. [27]

    Mmdetection: Open mmlab detection toolbox and benchmark,

    K. Chen, J. Wang, J. Pang, Y . Cao, Y . Xiong, X. Li, S. Sun, W. Feng, Z. Liu, J. Xuet al., “Mmdetection: Open mmlab detection toolbox and benchmark,”arXiv preprint arXiv:1906.07155, 2019

  20. [28]

    Yolo- world: Real-time open-vocabulary object detection,

    T. Cheng, L. Song, Y . Ge, W. Liu, X. Wang, and Y . Shan, “Yolo- world: Real-time open-vocabulary object detection,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 16 901–16 911

  21. [29]

    Cf-yolo: Cross fusion yolo for object detection in adverse weather with a high-quality real snow dataset,

    Q. Ding, P. Li, X. Yan, D. Shi, L. Liang, W. Wang, H. Xie, J. Li, and M. Wei, “Cf-yolo: Cross fusion yolo for object detection in adverse weather with a high-quality real snow dataset,”IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 10, pp. 10 749–10 759, 2023

  22. [30]

    Id-yolo: Real-time salient object detection based on the driver’s fixa- tion region,

    L. Qin, Y . Shi, Y . He, J. Zhang, X. Zhang, Y . Li, T. Deng, and H. Yan, “Id-yolo: Real-time salient object detection based on the driver’s fixa- tion region,”IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 9, pp. 15 898–15 908, 2022

  23. [31]

    Yolov7-rdd: A lightweight efficient pavement distress detection model,

    Z. Ning, H. Wang, S. Li, and Z. Xu, “Yolov7-rdd: A lightweight efficient pavement distress detection model,”IEEE Transactions on Intelligent Transportation Systems, 2024

  24. [32]

    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,” inComputer Vision– ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part I 14. Springer, 2016, pp. 21–37

  25. [33]

    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

  26. [34]

    Contextual object detection with multimodal large language models,

    Y . Zang, W. Li, J. Han, K. Zhou, and C. C. Loy, “Contextual object detection with multimodal large language models,”International Journal of Computer Vision, pp. 1–19, 2024

  27. [35]

    Taskclip: Extend large vision-language model for task oriented object detection,

    H. Chen, W. Huang, Y . Ni, S. Yun, Y . Liu, F. Wen, A. Velasquez, H. Lat- apie, and M. Imani, “Taskclip: Extend large vision-language model for task oriented object detection,”arXiv preprint arXiv:2403.08108, 2024

  28. [36]

    Deep learning-based crack damage detection using convolutional neural networks,

    Y .-J. Cha, W. Choi, and O. B ¨uy¨uk¨ozt¨urk, “Deep learning-based crack damage detection using convolutional neural networks,”Computer-Aided Civil and Infrastructure Engineering, vol. 32, no. 5, pp. 361–378, 2017

  29. [37]

    Structural damage detection with automatic feature-extraction through deep learning,

    Y .-z. Lin, Z.-h. Nie, and H.-w. Ma, “Structural damage detection with automatic feature-extraction through deep learning,”Computer-Aided Civil and Infrastructure Engineering, vol. 32, no. 12, pp. 1025–1046, 2017

  30. [38]

    Structural building damage detection with deep learning: Assessment of a state-of-the-art cnn in operational conditions,

    F. Nex, D. Duarte, F. G. Tonolo, and N. Kerle, “Structural building damage detection with deep learning: Assessment of a state-of-the-art cnn in operational conditions,”Remote sensing, vol. 11, no. 23, p. 2765, 2019

  31. [39]

    Skip connection yolo architecture for noise barrier defect detection using uav-based images in high-speed railway,

    J. Cui, Y . Qin, Y . Wu, C. Shao, and H. Yang, “Skip connection yolo architecture for noise barrier defect detection using uav-based images in high-speed railway,”IEEE Transactions on Intelligent Transportation Systems, 2023

  32. [40]

    Run, don’t walk: chasing higher flops for faster neural networks,

    J. Chen, S.-h. Kao, H. He, W. Zhuo, S. Wen, C.-H. Lee, and S.-H. G. Chan, “Run, don’t walk: chasing higher flops for faster neural networks,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 12 021–12 031

  33. [41]

    Efficient multi-scale attention module with cross-spatial learning,

    D. Ouyang, S. He, G. Zhang, M. Luo, H. Guo, J. Zhan, and Z. Huang, “Efficient multi-scale attention module with cross-spatial learning,” in ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2023, pp. 1–5

  34. [42]

    Inner-iou: more effective inter- section over union loss with auxiliary bounding box,

    H. Zhang, C. Xu, and S. Zhang, “Inner-iou: more effective inter- section over union loss with auxiliary bounding box,”arXiv preprint arXiv:2311.02877, 2023

  35. [43]

    Advanced auto labeling solution with added features,

    W. Wang, “Advanced auto labeling solution with added features,” 2023

  36. [44]

    Focal loss for dense object detection,

    T. Lin, “Focal loss for dense object detection,”arXiv preprint arXiv:1708.02002, 2017

  37. [45]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,”arXiv preprint arXiv:2103.14030, 2021

  38. [46]

    Lightweight rail surface defect detection algorithm based on an im- proved yolov8,

    X. CanYang, L. Yingying, L. Yongqiang, T. Runliang, and G. Tao, “Lightweight rail surface defect detection algorithm based on an im- proved yolov8,”Measurement, vol. 242, p. 115922, 2025

  39. [47]

    Lf-yolo: A lighter and faster yolo for weld defect detection of x-ray image,

    M. Liu, Y . Chen, J. Xie, L. He, and Y . Zhang, “Lf-yolo: A lighter and faster yolo for weld defect detection of x-ray image,”IEEE Sensors Journal, vol. 23, no. 7, pp. 7430–7439, 2023

  40. [48]

    Mcnet: Multiple context information segmentation network of no-service rail surface de- fects,

    D. Zhang, K. Song, J. Xu, Y . He, M. Niu, and Y . Yan, “Mcnet: Multiple context information segmentation network of no-service rail surface de- fects,”IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1–9, 2021

  41. [49]

    Dynamically weighted balanced loss: class imbalanced learning and confidence calibration of deep neural networks,

    K. R. M. Fernando and C. P. Tsokos, “Dynamically weighted balanced loss: class imbalanced learning and confidence calibration of deep neural networks,”IEEE Transactions on Neural Networks and Learning Systems, vol. 33, no. 7, pp. 2940–2951, 2021. JOURNAL OF LATEX CLASS FILES, ...

  42. [2023]

    His research focuses on Digital Modeling and Intelligent Equipment Design

    He is currently working toward his MSc degree at Nanjing Insititute of Technology (NJIT), Nanjing, China. His research focuses on Digital Modeling and Intelligent Equipment Design. Wei Zhaoreceived his Ph.D degree (2024) in Bi- ological Science & Medical Engineering from the S...

Pith tools

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