Attention-based vector quantisation variational autoencoder for colour-patterned fabrics defect detection / Hongwei Zhang in COLORATION TECHNOLOGY, Vol. 139, N° 3 (06/2023)  

[article]  inCOLORATION TECHNOLOGY > Vol. 139, N° 3 (06/2023) . - p. 223-238 Titre : Attention-based vector quantisation variational autoencoder for colour-patterned fabrics defect detection Type de document : texte imprimé Auteurs : Hongwei Zhang, Auteur ; Guanhua Qiao, Auteur ; Shuting Liu, Auteur ; Yuting Lyu, Auteur ; Le Yao, Auteur ; Zhiqiang Ge, Auteur Année de publication : 2023 Article en page(s) : p. 223-238 Note générale : Bibliogr. Langues : Anglais (eng) Catégories : Auto-encodeurs variationnels En apprentissage automatique, un auto-encodeur variationnel (ou VAE de l'anglais variational auto encoder), est une architecture de réseau de neurones artificiels introduite en 2013 par D. Kingma et M. Welling, appartenant aux familles des modèles graphiques probabilistes et des méthodes bayésiennes variationnelles. Les VAE sont souvent rapprochés des autoencodeurs en raison de leur architectures similaires. Leur utilisation et leur formulation mathématiques sont cependant différentes. Les auto-encodeurs variationnels permettent de formuler un problème d'inférence statistique (par exemple, déduire la valeur d'une variable aléatoire à partir d'une autre variable aléatoire) en un problème d'optimisation statistique (c'est-à-dire trouver les valeurs de paramètres qui minimisent une fonction objectif). Ils représentent une fonction associant à une valeur d'entrée une distribution latente multivariée, qui n'est pas directement observée mais déduite depuis un modèle mathématique à partir de la distribution d'autres variables. Bien que ce type de modèle ait été initialement conçu pour l'apprentissage non supervisé, son efficacité a été prouvée pour l'apprentissage semi-supervisé et l'apprentissage supervisé. (Wikipedia) Détection de défauts (Ingénierie) Textiles et tissus -- Défauts Index. décimale : 667.3 Teinture et impression des tissus Résumé : Defect detection is an essential link in the fabric production process. Due to the diversity of patterns and scarcity of defect samples for colour-patterned fabrics, reconstruction-based unsupervised deep learning algorithms have received extensive attention in the field of fabric defect detection. Among them, unsupervised reconstruction models based on variational autoencoders (VAEs) have been shown to be effective. However, there is a problem of posterior collapse in the process of modelling parametric distributions of continuous variables by VAEs. Therefore, VAE-based defect detection methods for colour-patterned fabrics usually produce ambiguous reconstruction results, thereby affecting the defect detection performance. In this article, an attention-based vector quantisation variational autoencoder (AVQ-VAE) is proposed for colour-patterned fabric defect detection. The method adopts autoregressive modelling of discrete variables to avoid the posterior collapse problem of traditional VAEs, and utilises attention mechanism to enhance the feature representation ability of the model. AVQ-VAE consists of encoder, embedding space, decoder and attention mechanism. The encoder is used to map the input image into multiple feature vectors. Vector quantisation in embedding space is used for discretisation and autoregressive modelling of feature vectors. A decoder is used to decode discrete variables into images of the same size as the original input. Furthermore, an attention mechanism is used to capture channel and spatial correlations, which help the model focus on important information by adaptively recalibrating feature maps. Experimental results on public datasets demonstrate that the proposed method is robust and effective for colour-patterned fabric defect detection. Note de contenu : - RELATED WORKS : Unsupervised deep learning based fabric defect detection - VAE-based model - Attention mechanism - THE PROPOSED AVQ-VAE METHOD : AVQ-VAE and training process - Colour-patterned fabric defect detection based on AVQ-VAE - EXPERIMENT : Experimental platform - Datasets - Parameter setting - Evaluation metrics - Defect detection results on dataset YDFID-1 - Defect detection results on anomaly detection dataset MVTec AD - Table 1 : Number of selected YDFID-1 patterns - Table 2 : Defect detection results of six models on colour-patterned fabrics - Table 3 : AUROC results for the three methodsTABLE 4. AUPRO results for the three methods DOI : https://doi.org/10.1111/cote.12644 En ligne : https://onlinelibrary.wiley.com/doi/epdf/10.1111/cote.12644 Format de la ressource électronique : Pdf Permalink : https://e-campus.itech.fr/pmb/opac_css/index.php?lvl=notice_display&id=39534 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
24086	-	Périodique	Bibliothèque principale	Documentaires	Disponible

Colour-patterned fabric defect detection based on an unsupervised multi-scale U-shaped denoising convolutional autoencoder model / Hongwei Zhang in COLORATION TECHNOLOGY, Vol. 138, N° 5 (10/2022)  

[article]  inCOLORATION TECHNOLOGY > Vol. 138, N° 5 (10/2022) . - p. 522-537 Titre : Colour-patterned fabric defect detection based on an unsupervised multi-scale U-shaped denoising convolutional autoencoder model Type de document : texte imprimé Auteurs : Hongwei Zhang, Auteur ; Shuting Liu, Auteur ; Quanlu Tan, Auteur ; Shuai Lu, Auteur ; Le Yao, Auteur ; Zhiqiang Ge, Auteur Année de publication : 2022 Article en page(s) : p. 522-537 Note générale : Bibliogr. Langues : Anglais (eng) Catégories : Détection de défauts (Ingénierie) Impression sur étoffes Textiles et tissus -- Défauts Index. décimale : 667.3 Teinture et impression des tissus Résumé : This study proposes an unsupervised, learning-based, reconstructed scheme and a residual analysis-based defect detection model for colour-patterned fabric defect detection problems in the clothing process industry. It solves the challenging problems of existing supervised fabric defect detection methods, such as high costs in manually labelling samples and designing features, unstable generalisation ability and scarcity of defective samples. First, for a specific texture, the training set was constructed by collecting easily accessible defect-free colour-patterned fabric images. Second, a multi-scale U-shaped denoising convolutional autoencoder was modelled using defect-free samples, which can reconstruct the newly tested colour-patterned fabric images automatically. Subsequently, a residual map between the original image and corresponding reconstructed image was calculated. Finally, the defective areas were detected and accurately localised by further opening operations. The experimental results indicated that the proposed method is valid and robust for detecting defects in various colour-patterned fabrics. Moreover, with the YDFID-1 dataset, compared with other models, the intersection over union index of the model proposed in the current paper was improved by at least 3.95%. Note de contenu : - PRELIMINARIES : The traditional U-Net mode - THE PROPOSED METHOD FOR COLOUR-PATTERNED FABRIC DEFECT DETECTION : The UDCAE model - The multi-scale UDCAE model - The model training phase - The model testing phase for defect detection - EXPERIMENTAL RESULTS AND ANALYSIS : Description of datasets - Experimental platform - Evaluation metrics - Qualitative and quantitative performance analyses - Table 1 : The number of colour-patterned fabric samples - Table 2 : The PSNR and SSIM values of reconstructed results with 11 loss functions - Table 3 : Comparison of time consumption of four loss functions in the training and testing phase - Table 4 : The PSNR and SSIM values of reconstructed defect-free results with four models - Table 5 : Comparison of four evaluation indicators of defect detection results with five models - Table 6 : Comparisons of AUROC of defect detection results with six models DOI : https://doi.org/10.1111/cote.12609 En ligne : https://onlinelibrary.wiley.com/doi/epdf/10.1111/cote.12609 Format de la ressource électronique : Pdf Permalink : https://e-campus.itech.fr/pmb/opac_css/index.php?lvl=notice_display&id=38126 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
23613	-	Périodique	Bibliothèque principale	Documentaires	Disponible

Colour-patterned fabric-defect detection using unsupervised and memorial defect-free features / Hongwei Zhang in COLORATION TECHNOLOGY, Vol. 138, N° 6 (12/2022)  

[article]  inCOLORATION TECHNOLOGY > Vol. 138, N° 6 (12/2022) . - p. 602-620 Titre : Colour-patterned fabric-defect detection using unsupervised and memorial defect-free features Type de document : texte imprimé Auteurs : Hongwei Zhang, Auteur ; Weiwei Zhang, Auteur ; Yang Wang, Auteur ; Shuai Lu, Auteur ; Le Yao, Auteur ; Xia Chen, Auteur Année de publication : 2022 Article en page(s) : p. 602-620 Note générale : Bibliogr. Langues : Anglais (eng) Catégories : Automatisation Détection de défauts (Ingénierie) Teinture -- Fibres textiles Textiles et tissus -- Défauts Textiles et tissus imprimés Index. décimale : 667.3 Teinture et impression des tissus Résumé : Automatic colour-patterned fabric-defect detection is essential and challenging in controlling manufacturing quality. Because of the scarcity of defective colour-patterned fabric samples and the imbalance of defect types, an auto-encoder trained with defect-free samples is used. However, the auto-encoder sometimes has weak generalisation ability, leading to mis- or over-detection of defects. An unsupervised and memorising defect-free method is proposed for colour-patterned fabric defects. The method designs a memory-guided quantisation variational auto-encoder-2 model and improves the residual post-processing operation. Specifically, it has three significant characteristics. First, it avoids time-consuming and laborious manually labelled samples and only needs defect-free samples in the training phase. Second, we expect the reconstruction image to be a defect-free image. Thus, memory modules are introduced to encourage the model to memorise the features of defect-free samples, which help to remove the image defect areas at the testing stage. Third, to further improve the detection accuracy, the closing operation is used to deal with the residual image that is calculated between the tested and the corresponding reconstructed image. Extensive experiments on various representative fabric samples demonstrated the effectiveness and superiority of the proposed method. Note de contenu : - RELATED WORK : Fabric-defect detection - Memory networks - METHOD: Network structure - Model training - Model testing - EXPERIMENTAL RESULTS AND ANALYSIS : Experimental platform - Experimental dataset - Evaluation indexes - Implementation details - Qualitative and quantitative analysis - Table 1 : The numbers of defective and non-defective samples in the YDFID-1 dataset - Table 2 : Comparison of detection performance of different threshold parameters and different mathematical morphology processing methods - Table 3 : PSNR and SSIM values of reconstruction results of seven models - Table 4 : Quantitative analysis and comparison of test results of seven models under four evaluation indexes DOI : https://doi.org/10.1111/cote.12624 En ligne : https://onlinelibrary.wiley.com/doi/epdf/10.1111/cote.12624 Format de la ressource électronique : Pdf Permalink : https://e-campus.itech.fr/pmb/opac_css/index.php?lvl=notice_display&id=38375 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
23716	-	Périodique	Bibliothèque principale	Documentaires	Disponible

Knowledge distillation for unsupervised defect detection of yarn-dyed fabric using the system DAERD : Dual attention embedded reconstruction distillation / Hongwei Zhang in COLORATION TECHNOLOGY, Vol. 140, N° 1 (02/2024)  

[article]  inCOLORATION TECHNOLOGY > Vol. 140, N° 1 (02/2024) . - p. 125-143 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 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
24413	-	Périodique	Bibliothèque principale	Documentaires	Disponible

A mixed-attention-based multi-scale autoencoder algorithm for fabric defect detection / Hongwei Zhang in COLORATION TECHNOLOGY, Vol. 140, N° 3 (06/2024)  

[article]  inCOLORATION TECHNOLOGY > Vol. 140, N° 3 (06/2024) . - p. 451-466 Titre : A mixed-attention-based multi-scale autoencoder algorithm for fabric defect detection Type de document : texte imprimé Auteurs : Hongwei Zhang, Auteur ; Yanzi Wu, Auteur ; Shuai Lu, Auteur ; Le Yao, Auteur ; Pengfei Li, Auteur Année de publication : 2024 Article en page(s) : p. 451-466 Note générale : Bibliogr. Langues : Anglais (eng) Catégories : Algorithmes Analyse multiéchelle Détection de défauts (Ingénierie) Textiles et tissus -- Défauts Résumé : Aiming at the defects in the process of fabric production, a defect detection model of fabric based on a mixed-attention-based multi-scale non-skipping U-shaped deep convolutional autoencoder (MADCAE) was proposed. In a traditional encoder, the convolutional layer treats each pixel equally, so the importance of different pixels cannot be reflected. It is difficult to obtain richer and more effective information. The reconstruction of the defect region and the detection of the defect region are further affected. In this article, three different scale features of input images are extracted by enlarging the receptive field with large kernel convolution blocks. A hybrid attention module is used to ensure the richness of the extracted information in terms of space and channel. Experiments show that this method can effectively reconstruct fabric parts without requiring a large number of defect marking samples. It can quickly detect and locate defective areas of fabric patterns. Note de contenu : - RELATED WORKS : Fabric defect detection method based on deep learning - Attention mechanism - M-Net - METHOD : Proposed model - Training of MADCAE-based models - Defect detection of yarn-dyed fabric parts based on the MADCAE model - EXPERIMENT : Dataset - Introduction to the experimental platform - Evaluation indicators - Analysis of experimental results - Ablation experiments DOI : https://doi.org/10.1111/cote.12725 En ligne : https://drive.google.com/file/d/13UA30D1lBJDI0x_rBRWtJcaRwHh-eJmg/view?usp=drive [...] Format de la ressource électronique : Pdf Permalink : https://e-campus.itech.fr/pmb/opac_css/index.php?lvl=notice_display&id=40962 [article]

Réservation

Réserver ce document

Exemplaires (1)

Code-barres	Cote	Support	Localisation	Section	Disponibilité
24657	-	Périodique	Bibliothèque principale	Documentaires	Disponible

---

1 (1 - 5 / 5)