| Titre : |
Knowledge distillation for unsupervised defect detection of yarn-dyed fabric using the system DAERD : Dual attention embedded reconstruction distillation |
| Type de document : |
texte imprimé |
| Auteurs : |
Hongwei Zhang, Auteur ; Shuaibo Liu, Auteur ; Shuai Lu, Auteur ; Le Yao, Auteur ; Pengfei Li, Auteur |
| Année de publication : |
2024 |
| Article en page(s) : |
p. 125-143 |
| Note générale : |
Bibliogr. |
| Langues : |
Anglais (eng) |
| Catégories : |
Détection de défauts (Ingénierie)
Textiles et tissus teints
|
| Index. décimale : |
667.3 Teinture et impression des tissus |
| Résumé : |
Detecting defects of yarn-dyed fabrics automatically in industrial scenarios can improve economic efficiency, but the scarcity of defect samples makes the task more challenging in the customised and small-batch production scenario. At present, most reconstruction-based methods have high requirements on the effect of reconstructing the defect area into the normal area, and the reconstruction performance often determines the final defect detection result. To solve this problem, this article proposes an unsupervised learning framework of dual attention embedded reconstruction distillation. We try to use this novel distillation scheme to provide some contribution to the defect detection field. Firstly, different from the encoder-encoder structure of traditional distillation, the teacher-student network in this article adopts the encoder-decoder structure. The purpose of the student network is to restore the normal feature representation of the pre-trained teacher network. Secondly, this article proposes a dual attention residual module, which can effectively remove redundant information and defective feature information from the teacher network through the double feature weight allocation mechanism. This helps the student network to recover the normal feature information output by the teacher network. Finally, the multi-level training deployment at the feature level in this article aims to make the model obtain accurate defect detection results. The proposed method has been extensively tested on the published fabric dataset YDFID-1, ZJU-Leaper dataset and the anomaly detection dataset MVTec. The results show that this method not only has good performance in fabric defect detection and location but also has universal applicability. |
| Note de contenu : |
- RELATED WORKS : Fabric defect detection - Attention mechanism
- METHODS : The proposed DAERD model - DAERD-based yarn-dyed fabric defect detection
- EXPERIMENT : Experimental details - Datasets - Evaluation metric - Experimental results and comparative analysis
- Table 1 : Distribution of the number of defective and defect-free samples in the partial dataset of YDFID-1
- Table 2 : Distribution of the number of anomaly-free and anomalous samples in the texture class dataset
- Table 3 : Image-level AUROC (%) of defect detection methods on the YDFID-1 dataset
- Table 4 : Pixel-level AUROC (%) and AUPRO (%) of defect localisation methods on the YDFID-1 dataset
- Table 5 : Image-level AUROC (%) of defect detection methods on the MVTec dataset
- Table 6 : Pixel-level AUROC (%) and AUPRO (%) of defect localisation methods on the MVTec dataset
- Table 7 : Image-level AUROC (%) of defect detection methods on the ZJU-Leaper dataset
- Table 8 : Pixel-level AUROC (%) and AUPRO (%) of defect localisation methods on the ZJU-Leaper dataset
- Table 9 : Ablation study of the effect of SimAM (simple, parameter-free, attention module) and SE (squeeze-and-excitation) network on the model performance |
| DOI : |
https://doi.org/10.1111/cote.12705 |
| En ligne : |
https://drive.google.com/file/d/1fzssAGVLxuRnKx8FtikkhnIF1yv1cue3/view?usp=drive [...] |
| Format de la ressource électronique : |
Pdf |
| Permalink : |
https://e-campus.itech.fr/pmb/opac_css/index.php?lvl=notice_display&id=40396 |
in COLORATION TECHNOLOGY > Vol. 140, N° 1 (02/2024) . - p. 125-143
Aller sur edunet
Accueil
Knowledge distillation for unsupervised defect detection of yarn-dyed fabric using the system DAERD : Dual attention embedded reconstruction distillation in COLORATION TECHNOLOGY, Vol. 140, N° 1 (02/2024)
