Woven Fabric Defect Detection Based on Convolutional Neural Network for Binary Classification

Gao, Chengde; Zhou, Jie; Wong, Wai Keung; Gao, Tianyu

doi:10.1007/978-3-319-99695-0_37

Cited by 16 publications

(11 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This leads to a binary classification problem. In Gao et al (2018), a defect detection accuracy of 96.5% has been achieved. The only one work dealing with the classification of a more abstract fabric property is our own previous work (Dorozynski et al, 2019), where we used a network on top of a pre-trained ResNet-152 to predict the time of production of silk fabrics.…”

Section: Related Workmentioning

confidence: 97%

See 1 more Smart Citation

Multi-Task Deep Learning With Incomplete Training Samples for The Image-Based Prediction of Variables Describing Silk Fabrics

Dorozynski

Clermont

Rottensteiner

2019

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

<p><strong>Abstract.</strong> This paper presents a method for the classification of images of silk fabrics with the aim to predict properties such as the place and time of origin and the production technique. The proposed method was developed in the context of the EU project SILKNOW (<a href="http://silknow.eu/"target="_blank">http://silknow.eu/</a>). In the context of classification, we address the problem of limited as well as not fully labelled data and investigate the connection between the distinct variables. A pre-trained Convolutional Neural Network (CNN) is used for the feature extraction and a classification network realizing Multi-task learning (MTL) is trained based on these features. The training procedure is adapted to enable the consideration of images that do not have a label for all tasks. Additionally, MTL with fully labeled training data is investigated for the classification of silk fabrics. The impact of both MTL approaches is compared to singletask learning based on two different class structures. We achieve overall accuracies of 92&ndash;95&thinsp;% and average F1-scores of 88&ndash;90&thinsp;% in our best experiments. </p>

show abstract

Section: Related Workmentioning

confidence: 97%

“…A CNNbased classifier was trained to predict different patterns of knitted fabrics in Xiao et al (2018), achieving an overall accuracy of 98.4% among eight categories of structures. There are also contributions to detect fabric defects based on images (Gao et al, 2018). This leads to a binary classification problem.…”

Section: Related Workmentioning

confidence: 99%

Multi-Task Deep Learning With Incomplete Training Samples for The Image-Based Prediction of Variables Describing Silk Fabrics

Dorozynski

Clermont

Rottensteiner

2019

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…Whereas all these works investigated the derivation of one single characteristic of images of fabrics, Meng and others [75] proposed a multi-task neural network to simultaneously locate individual yarns in a fabric and determine a float-point location map. To improve the available knowledge about objects in a digital heritage collection, e.g., fabrics, one cannot only predict the properties that are directly related to the visual appearance of a fabric (e.g., the pattern of a fabric or fabric defects [76,77], but also more abstract ones). Li and others [78] assessed this task for cultural assets by means of hierarchically classifying first the type of asset, e.g., a painting, followed by a type specific classification, e.g., the artist of a painting.…”

Section: Multi-task Learning With Training Samples For the Image-basementioning

confidence: 99%

From Silk to Digital Technologies: A Gateway to New Opportunities for Creative Industries, Traditional Crafts and Designers. The SILKNOW Case

et al. 2020

View full text Add to dashboard Cite

Nowadays, cultural heritage is more than ever linked to the present. It links us to our cultural past through the conscious act of preserving and bequeathing to future generations, turning society into its custodian. The appreciation of cultural heritage happens not only because of its communicative power, but also because of its economic power, through sustainable development and the promotion of creative industries. This paper presents SILKNOW, an EU-H2002 funded project and its application to cultural heritage, as well as to creative industries and design innovation. To this end, it presents the use of image recognition tools applied to cultural heritage, through the interoperability of data in the open-access registers of silk museums and its presentation, analysis and creative process carried out by the design students of EASD Valencia as a case study, in the branches of jewellery and fashion project, inspired by the heritage of silk.

show abstract

“…For fabrics of high complexity (i.e., multimodal appearance), supervised approaches that require both normal and anomalous data [6,7] are predominantly used. For example, classification, segmentation and object detection approaches have been successfully adapted to the fabric inspection task [12][13][14][15][16][17]. Moreover, supervised algorithms generally outperform their semi-supervised counterparts [18,19].…”

Section: Introductionmentioning

confidence: 99%

Increasing the Generalization of Supervised Fabric Anomaly Detection Methods to Unseen Fabrics

Rippel

Zwinge

Merhof

2022

Sensors

View full text Add to dashboard Cite

Fabric ad tries to detect anomalies (i.e., defects) in fabrics, and fabric ad approaches are continuously improved with respect to their ad performance. However, developed solutions are known to generalize poorly to previously unseen fabrics, posing a crucial limitation to their applicability. Moreover, current research focuses on adapting converged models to previously unseen fabrics in a post hoc manner, rather than training models that generalize better in the first place. In our work, we explore this potential for the first time. Specifically, we propose that previously unseen fabrics can be regarded as shifts in the underlying data distribution. We therefore argue that factors which reportedly improve a model’s resistance to distribution shifts should also improve the performance of supervised fabric ad methods on unseen fabrics. Hence, we assess the potential benefits of: (I) vrm techniques adapted to the fabric ad use-case, (II) different loss functions, (III) ImageNet pre-training, (IV) dataset diversity, and (V) model architecture as well as model complexity. The subsequently performed large-scale analysis reveals that (I) only the vrm technique, AugMix, consistently improves performance on unseen fabrics; (II) hsc outperforms other loss functions when combined with AugMix and (III) ImageNet pre-training, which is already beneficial on its own; (IV) increasing dataset diversity improves performance on unseen fabrics; and (V) architectures with better ImageNet performance also perform better on unseen fabrics, yet the same does not hold for more complex models. Notably, the results show that not all factors and techniques which reportedly improve a model’s resistance to distribution shifts in natural images also improve the generalization of supervised fabric ad methods to unseen fabrics, demonstrating the necessity of our work. Additionally, we also assess whether the performance gains of models which generalize better propagate to post hoc adaptation methods and show this to be the case. Since no suitable fabric dataset was publicly available at the time of this work, we acquired our own fabric dataset, called OLP, as the basis for the above experiments. OLP consists of 38 complex, patterned fabrics, more than 6400 images in total, and is made publicly available.

show abstract

Woven Fabric Defect Detection Based on Convolutional Neural Network for Binary Classification

Cited by 16 publications

References 5 publications

Multi-Task Deep Learning With Incomplete Training Samples for The Image-Based Prediction of Variables Describing Silk Fabrics

Multi-Task Deep Learning With Incomplete Training Samples for The Image-Based Prediction of Variables Describing Silk Fabrics

From Silk to Digital Technologies: A Gateway to New Opportunities for Creative Industries, Traditional Crafts and Designers. The SILKNOW Case

Increasing the Generalization of Supervised Fabric Anomaly Detection Methods to Unseen Fabrics

Contact Info

Product

Resources

About