DEFECT ENHANCEMENT DETECTION METHOD IN COMPLEX BACKGROUNDS BASED ON MIXED ATTENTION

Authors

DOI:

https://doi.org/10.35631/JISTM.1142002

Keywords:

Attention Mechanism, Feature Enhancement, Defect Detection, Mixed Attention, Lightweight Network

Abstract

This paper proposes a defect enhancement detection method based on Mixed attention, aiming to address the difficulty of feature extraction caused by complex background interference in metal surface defect detection. The core of this method is the design of a lightweight Mixed Attention Module (MAM). This module integrates channel attention and spatial attention mechanisms in parallel, working collaboratively at multiple levels: channel attention adaptively recalibrates channel feature responses by modeling the interdependencies between feature channels, emphasizing feature maps related to defects; spatial attention focuses on key spatial locations in the feature maps, generating spatial weight masks to highlight defect regions and suppress texture and noise interference from irrelevant backgrounds. Simultaneously, the module employs an efficient structural design, achieving effective capture and fusion of multi-scale contextual information without introducing significant computational overhead, thereby enhancing the discriminative representation of defect features in complex backgrounds. Experimental results demonstrate that this method achieves significant improvements in detection accuracy (mAP) on the NEU-DET and GDUT-DET public metal surface defect datasets.

Downloads

Download data is not yet available.

References

Akhyar, F., Furqon, E. N., & Lin, C.-Y. (2022). Enhancing precision with an ensemble generative adversarial network for steel surface defect detectors (EnsGAN-SDD). Sensors, 22(11), 4257. https://doi.org/10.3390/s22114257

Du, W., Shen, H., Fu, J., Zhang, G., & He, Q. (2019). Approaches for improvement of the X-ray image defect detection of automobile casting aluminum parts based on deep learning. NDT & E International, 107, 102144. https://doi.org/10.1016/j.ndteint.2019.102144

Ge, Z., Liu, S., Wang, F., Li, Z., & Sun, J. (2021). YOLOX: Exceeding YOLO series in 2021. arXiv. https://doi.org/10.48550/arXiv.2107.08430

Hussain, M. (2023). YOLO-v1 to YOLO-v8, the rise of YOLO and its complementary nature toward digital manufacturing and industrial defect detection. Machines, 11(7), 677. https://doi.org/10.3390/machines11070677

Jeong, J., Park, H., & Kwak, N. (2017). Enhancement of SSD by concatenating feature maps for object detection. arXiv. https://doi.org/10.48550/arXiv.1705.09587

Li, C., Li, L., Jiang, H., Weng, K., Geng, Y., Li, L., Ke, Z., Li, Q., Cheng, M., Nie, W., Li, Y., Zhang, B., Liang, Y., Zhou, L., Xu, X., Chu, X., Wei, X., & Wei, X. (2022). YOLOv6: A single-stage object detection framework for industrial applications. arXiv. https://doi.org/10.48550/arXiv.2209.02976

Liang, F., Zhao, L., Ren, Y., Wang, S., To, S., Abbas, Z., & Islam, M. S. (2024). LAD-Net: A lightweight welding defect surface non-destructive detection algorithm based on the attention mechanism. Computers in Industry, 161, 104109. https://doi.org/10.1016/j.compind.2024.104109

Ma, J., Hu, S., Fu, J., & Chen, G. (2024). A hierarchical attention detector for bearing surface defect detection. Expert Systems with Applications, 239, 122365. https://doi.org/10.1016/j.eswa.2023.122365

Ma, L., Xie, W., & Zhang, Y. (2019). Blister defect detection based on convolutional neural network for polymer lithium-ion battery. Applied Sciences, 9(6), 1085. https://doi.org/10.3390/app9061085

Redmon, J., & Farhadi, A. (2018). YOLOv3: An incremental improvement. arXiv. https://doi.org/10.48550/arXiv.1804.02767

Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2017). Inception-v4, Inception-ResNet and the impact of residual connections on learning. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1), 4278–4284. https://doi.org/10.1609/aaai.v31i1.11231

Wang, X., Ma, S., Wu, S., Li, Z., Cao, J., & Xu, P. (2025). Detection of surface defects in steel based on dual-backbone network: MBDNet-attention-YOLO. Sensors, 25(15), 4817. https://doi.org/10.3390/s25154817

Wang, Z., & Liu, W. (2024). Surface defect detection algorithm for strip steel based on improved YOLOv7 model. IAENG International Journal of Computer Science, 51(3), 308–317.

Zhao, C., Shu, X., Yan, X., Zuo, X., & Zhu, F. (2023). RDD-YOLO: A modified YOLO for detection of steel surface defects. Measurement, 214, 112776. https://doi.org/10.1016/j.measurement.2023.112776

Zong, R., Liu, Q., Wang, J., Jia, X., Qin, N., & Huang, D. (2024). A metal surface defect detection algorithm based on mixed supervised and cross stage partial darknet. In 2024 43rd Chinese Control Conference (CCC) (pp. 8106–8111). https://doi.org/10.23919/CCC63176.2024.10661652

Downloads

Published

2026-02-22

How to Cite

Changyin , C., & Ahmad, M. N. (2026). DEFECT ENHANCEMENT DETECTION METHOD IN COMPLEX BACKGROUNDS BASED ON MIXED ATTENTION. JOURNAL INFORMATION AND TECHNOLOGY MANAGEMENT (JISTM), 11(42), 16–31. https://doi.org/10.35631/JISTM.1142002