Home > Research > Publications & Outputs > Dual-Metric Neural Network with Attention Guida...
View graph of relations

Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing

Research output: Contribution to Journal/MagazineJournal articlepeer-review

E-pub ahead of print

Standard

Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing. / Gao, Pengjie; Wang, Junliang; Xia, Min et al.
In: Journal of Manufacturing Science and Engineering, 07.09.2023, p. 1-31.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Gao, P, Wang, J, Xia, M, Qin, Z & Zhang, J 2023, 'Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing', Journal of Manufacturing Science and Engineering, pp. 1-31. https://doi.org/10.1115/1.4063356

APA

Gao, P., Wang, J., Xia, M., Qin, Z., & Zhang, J. (2023). Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing. Journal of Manufacturing Science and Engineering, 1-31. Advance online publication. https://doi.org/10.1115/1.4063356

Vancouver

Gao P, Wang J, Xia M, Qin Z, Zhang J. Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing. Journal of Manufacturing Science and Engineering. 2023 Sept 7;1-31. Epub 2023 Sept 7. doi: 10.1115/1.4063356

Author

Gao, Pengjie ; Wang, Junliang ; Xia, Min et al. / Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing. In: Journal of Manufacturing Science and Engineering. 2023 ; pp. 1-31.

Bibtex

@article{fc641babd8c04715806e6239222f0348,
title = "Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing",
abstract = "As an important application of human-robot collaboration, intelligent detection of surface defects is crucial for production quality control, which also helps in relieving the workload of technical staff in human-centric smart manufacturing. To accurately detect defects with limited samples in industrial practice, a dual-metric neural network with attention guided is proposed. First, an attention-guided recognition network with channel attention and position attention module is designed to efficiently learn representative defect features with limited samples. Second, aiming to detect defects with confusing surface images, a dual-metric function is presented to learn the classification boundary by controlling the distance of samples in feature space from intra-class and inter-class. The experiment results on the fabric defect dataset demonstrate that the proposed approach outperforms state-of-the-art methods in accuracy, recall, precision, F1-score, and few-shot accuracy. Further comparative experiments reveal that the dual-metric function is superior in improving the few-shot detection accuracy for the defect patterns of fabric.",
keywords = "Industrial and Manufacturing Engineering, Computer Science Applications, Mechanical Engineering, Control and Systems Engineering",
author = "Pengjie Gao and Junliang Wang and Min Xia and Zijin Qin and Jie Zhang",
year = "2023",
month = sep,
day = "7",
doi = "10.1115/1.4063356",
language = "English",
pages = "1--31",
journal = "Journal of Manufacturing Science and Engineering",
issn = "1087-1357",
publisher = "American Society of Mechanical Engineers(ASME)",

}

RIS

TY - JOUR

T1 - Dual-Metric Neural Network with Attention Guidance for Surface Defect Few-Shot Detection in Smart Manufacturing

AU - Gao, Pengjie

AU - Wang, Junliang

AU - Xia, Min

AU - Qin, Zijin

AU - Zhang, Jie

PY - 2023/9/7

Y1 - 2023/9/7

N2 - As an important application of human-robot collaboration, intelligent detection of surface defects is crucial for production quality control, which also helps in relieving the workload of technical staff in human-centric smart manufacturing. To accurately detect defects with limited samples in industrial practice, a dual-metric neural network with attention guided is proposed. First, an attention-guided recognition network with channel attention and position attention module is designed to efficiently learn representative defect features with limited samples. Second, aiming to detect defects with confusing surface images, a dual-metric function is presented to learn the classification boundary by controlling the distance of samples in feature space from intra-class and inter-class. The experiment results on the fabric defect dataset demonstrate that the proposed approach outperforms state-of-the-art methods in accuracy, recall, precision, F1-score, and few-shot accuracy. Further comparative experiments reveal that the dual-metric function is superior in improving the few-shot detection accuracy for the defect patterns of fabric.

AB - As an important application of human-robot collaboration, intelligent detection of surface defects is crucial for production quality control, which also helps in relieving the workload of technical staff in human-centric smart manufacturing. To accurately detect defects with limited samples in industrial practice, a dual-metric neural network with attention guided is proposed. First, an attention-guided recognition network with channel attention and position attention module is designed to efficiently learn representative defect features with limited samples. Second, aiming to detect defects with confusing surface images, a dual-metric function is presented to learn the classification boundary by controlling the distance of samples in feature space from intra-class and inter-class. The experiment results on the fabric defect dataset demonstrate that the proposed approach outperforms state-of-the-art methods in accuracy, recall, precision, F1-score, and few-shot accuracy. Further comparative experiments reveal that the dual-metric function is superior in improving the few-shot detection accuracy for the defect patterns of fabric.

KW - Industrial and Manufacturing Engineering

KW - Computer Science Applications

KW - Mechanical Engineering

KW - Control and Systems Engineering

U2 - 10.1115/1.4063356

DO - 10.1115/1.4063356

M3 - Journal article

SP - 1

EP - 31

JO - Journal of Manufacturing Science and Engineering

JF - Journal of Manufacturing Science and Engineering

SN - 1087-1357

ER -