journal article·2026-07-30

YOLO11-DSU: An enhanced YOLO11 framework for defect detection in dark woven fabrics

Guanghui Xu, Weizhen Wang, Yuan Fang

Publisher page

Abstract

Although computer vision technologies have progressed rapidly, leading detectors experience severe accuracy degradation on low-contrast dark fabrics, frequently yielding mean average precision (mAP) scores below 50%. To address these inherent limitations, this work introduces YOLO11-DSU, an enhanced object identification paradigm constructed upon the foundational YOLO11 baseline. To counteract representation degradation encountered within poorly illuminated environments, a novel dual-backbone structure is integrated into the architecture. Furthermore, a space-channel synergistic attention (SCSA) mechanism is embedded at the fusion node, forming an end-to-end optimization process regulated by the adaptive unified intersection over union (UIoU) loss function. This configuration guarantees the deep integration of multilevel features while effectively mitigating noise. Empirical evaluations indicate that the mAP@0.5 of YOLO11-DSU reaches 80.3% on the TILDA dataset and 48.3% on the Tianchi dataset, outperforming the baseline by absolute margins of 6.8 and 4.1 percentage points, respectively. The proposed work establishes a robust technical framework for automated quality testing in high-end textile manufacturing.
Source: crossref · Quality: baseline

Affiliations

Clothing Human Factors and Artificial Intelligence Design Research Center, Dalian Polytechnic University, Dalian, China

School of Innovation and Entrepreneurship, Dalian Polytechnic University, Dalian, China

Funding

the Teaching Reform Project of the Department of Education of Liaoning Province · scientific research fund of liaoning provincial education department · the ‘Textile Light’ Teaching Reform Project of the China National Textile and Apparel Council · Liaoning Provincial Education Science Planning Project

References (39)

Show source-provided references
1. 10.1007/s40747-023-01317-810.1007/s40747-023-01317-8 ↗
2. Carvalho A, Miguéis V, Moura e Sá M. Enhancing operational performance in textile manufacturing: impact of deep learning-based defect detection. Int J Prod Res 2026: 1-25.
3. 10.1016/j.eswa.2025.12726110.1016/j.eswa.2025.127261 ↗
4. 10.1109/TII.2024.350795410.1109/TII.2024.3507954 ↗
5. 10.1109/TIM.2025.355882810.1109/TIM.2025.3558828 ↗
6. 10.1109/ACCESS.2024.342928310.1109/ACCESS.2024.3429283 ↗
7. 10.3390/s2507227010.3390/s25072270 ↗
8. Wang J, Cheng J, Gao C, et al. Enhanced fabric defect detection with feature contrast interference suppression. IEEE Trans Instrum Meas 2025; 74: 1-12.
9. 10.1038/s41598-024-57990-310.1038/s41598-024-57990-3 ↗
10. 10.1109/TASE.2024.335869310.1109/TASE.2024.3358693 ↗
11. 10.3390/app1506322810.3390/app15063228 ↗
12. 10.1109/ACCESS.2025.357045510.1109/ACCESS.2025.3570455 ↗
13. 10.1109/ACCESS.2025.352877110.1109/ACCESS.2025.3528771 ↗
14. 10.1080/15440478.2024.240152510.1080/15440478.2024.2401525 ↗
15. 10.1016/j.eswa.2025.12653710.1016/j.eswa.2025.126537 ↗
16. 10.7717/peerj-cs.313610.7717/peerj-cs.3136 ↗
17. 10.3390/app1401004110.3390/app14010041 ↗
18. 10.1016/j.neucom.2025.12986610.1016/j.neucom.2025.129866 ↗
19. 10.1016/j.conbuildmat.2024.13502510.1016/j.conbuildmat.2024.135025 ↗
20. Tong Z Chen Y Xu Z et al. Wise-IoU: bounding box regression loss with dynamic focusing mechanism. arXiv Preprint 2023; arXiv:2301.10051.
21. Luo X Cai Z Shao B et al. Unified-IoU: For high-quality object detection. arXiv Preprint 2024; arXiv:2408.06636.
22. 10.1371/journal.pone.032256110.1371/journal.pone.0322561 ↗
23. Mei S, Xiao Z, Chen J, et al. Improved YOLOv5s fabric defect detection algorithm based on the fusion of global attention mechanism and decoupling head. Int J Inf Retr Res 2024; 14(1): 1-18.
24. 10.1109/TCYB.2021.309530510.1109/TCYB.2021.3095305 ↗
25. 10.1109/TNNLS.2025.356236310.1109/TNNLS.2025.3562363 ↗
26. 10.1109/TII.2025.352856110.1109/TII.2025.3528561 ↗
27. 10.1302/2046-3758.149.BJR-2024-0505.R110.1302/2046-3758.149.BJR-2024-0505.R1 ↗
28. Redmon J Farhadi A. YOLOv3: An incremental improvement. arXiv Preprint 2018; arXiv:1804.02767.
29. Li C Li L Jiang H et al. YOLOv6: A single-stage object detection framework for industrial applications. arXiv Preprint 2022; arXiv:2209.02976.
30. 10.1109/ADICS58448.2024.1053361910.1109/ADICS58448.2024.10533619 ↗
31. Wang A, Chen H, Liu L, et al. YOLOv10: Real-time end-to-end object detection. In: Proceedings of the Conference on Neural Information Processing Systems (NeurIPS), 2024, pp. 107984–108011.
32. Alif MAR Hussain M. YOLOv12: A breakdown of the key architectural features. arXiv Preprint 2025; arXiv:2502.14740.
33. Sapkota R Cheppally RH Sharda A et al. YOLO26: Key architectural enhancements and performance benchmarking for real-time object detection. arXiv Preprint 2025; arXiv:2509.25164.
34. 10.1177/0040517525135619110.1177/00405175251356191 ↗
35. 10.1177/0040517525131506910.1177/00405175251315069 ↗
36. Hou Q, Zhou D, Feng J, et al. Coordinate attention for efficient mobile network design. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 13713–13722.
37. Woo S, Park J, Lee JY, et al. CBAM: Convolutional block attention module. In: Proceedings of the European Conference on Computer Vision. Berlin: Springer, 2018, pp. 3–19.
38. 10.1109/CVPR42600.2020.0115510.1109/CVPR42600.2020.01155 ↗
39. Liu Y Shao Z Hoffmann N et al. Global attention mechanism: Retain information to enhance channel-spatial interactions. arXiv Preprint 2021; arXiv:2112.05561.