Automatic fabric defect detection using a deep convolutional neural network

Jing, Junfeng; Ma, Hao; Zhang, Huan‐Huan

doi:10.1111/cote.12394

Cited by 125 publications

(71 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This section describes a set of experiments to evaluate the performance of the proposed method. The proposed method is compared with two fabric defect detection methods, PTIP [20] and LGM-FC [13], in terms of detection speed and accuracy. Accurately, to illustrate the detection speed of the proposed model, a comparison was made with several related methods in terms of detection time and the number of model parameters.…”

Section: Experiments and Discussionmentioning

confidence: 99%

“…Ouyang [19] proposed a fabric defect segmentation method based on convolution neural network embedded in an active layer. In order to detect pattern fabric defects, a hybrid method of traditional image processing and deep learning is proposed [20], which can achieve accurate detection of common defects in yarn-dyed fabric, such as holes, carrying, knots. Although the above methods use deep learning to extract features and achieve excellent detection performance automatically, they are all supervised learning methods, which need to collect, clean, and label training data sets.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Pixel-Wise Fabric Defect Detection by CNNs Without Labeled Training Data

Wang

Jing

2020

IEEE Access

View full text Add to dashboard Cite

Surface inspection is a necessary process of fabric quality control. However, it remains a challenging task owing to diverse types of defects, various patterns of fabric texture, and application requirements for detection speed. In this paper, a lightweight deep learning model is therefore proposed to complete the segmentation of fabric defects. The input of the model is a fabric image, and the output is a binary image. Generally known, a deep learning model usually needs much data to update the parameters. Still, as an abnormal phenomenon, fabric defects are unpredictable, which makes it impossible to collect a large number of data. Distinct from other models, the proposed method is a supervised network but does not need manually labeled samples for training. A fake sample generator is designed to simulate the defect image, which only needs the defect-free fabric image. The proposed model is trained with fake samples and verified with real samples. The experimental results show that the model trained with false data is useful and achieves high segmentation accuracy on real fabric samples. Besides, a loss function is proposed to deal with the problem of imbalance between the number of background pixels and the number of defective pixels in the fabric image. Comprehensive experiments were performed on representative fabric samples to verify the segmentation accuracy and detection speed of this method. INDEX TERMS Fabric defect; deep learning; image segmentation; defect detection; imbalanced dataset.

show abstract

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Pixel-Wise Fabric Defect Detection by CNNs Without Labeled Training Data

Wang

Jing

2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…This technique was highly efficient in fabric defect detection, with accuracy above 94 percent as presented in research work [10]. In other publications, defects, especially regarding colour disturbances were investigated by Jing et al [11]. The convolutional neural network (CNN) was involved in this evaluation.…”

Section: Introductionmentioning

confidence: 96%

“…The mentioned fabrics were divided by weaving pattern and technological parameters such a warp and weft density and mass per unit area. Fabrics 4 to 6 represented different abrasion resistance values, tested according to the standard [11], from 10,000 up to 25,000 rubs. The pilling phenomena occurred during the tests.…”

mentioning

confidence: 99%

The Algorithms of Image Processing and Analysis in the Textile Fabrics Abrasion Assessment

Jasińska

2019

Applied Sciences

View full text Add to dashboard Cite

The abrasion resistance and susceptibility of textiles to surface damage is an important issue for their utility properties. The currently used test method for abrasion resistance assessment has been based on criteria that define the breakage point of fabrics, visually identified and highly depended on observer perception. Taking under consideration the character of abrasion assessment process the efforts were made to elaborate the new alternative test technique which supports this evaluation. The abrasion resistance tests using instrumental method (based on image analysis tools) parallel with standard method were carried out. There were two steps of analysis. Firstly, the preliminary analysis for selecting the most effective and sensitive algorithm was carried out. Secondly, the final analysis, involving whole range of captured image samples, using chosen algorithm, based on image histogram calculation was done. Summarizing the abrasion resistance tests results obtained using both standard and instrumental methods it was found that the instrumental method gave more complex results during identification of surface changes caused directly by the abrasion process. The instrumental method is more sensitive to surface texture modifications and colour fade and its discrimination threshold is significantly lower than standard methods with qualitative breakage criteria (such as broken threads, loss of pile).

show abstract

“…The model employs Cascade AutoEncoder (CAE) to reconstruct the defects, and then these defects are segmented through a sharpening process, which is based on the assumption that the reconstructed images contain only normal features. Jing et al [21] presented a fabric anomaly inspection algorithm, which can detect various kinds of fabric defects. It's worth mentioning that it not only directly utilizes the original images as input, but it also divides the fabric images into multiple patches along the natural cycle of the fabric surface as the operation objects to train a CNN model.…”

Section: Introductionmentioning

confidence: 99%

A One-Stage Approach for Surface Anomaly Detection with Background Suppression Strategies

Liu

Yang

Guo

et al. 2020

Sensors

View full text Add to dashboard Cite

We explore a one-stage method for surface anomaly detection in industrial scenarios. On one side, encoder-decoder segmentation network is constructed to capture small targets as much as possible, and then dual background suppression mechanisms are designed to reduce noise patterns in coarse and fine manners. On the other hand, a classification module without learning parameters is built to reduce information loss in small targets due to the inexistence of successive down-sampling processes. Experimental results demonstrate that our one-stage detector achieves state-of-the-art performance in terms of precision, recall and f-score.

show abstract

Automatic fabric defect detection using a deep convolutional neural network

Cited by 125 publications

References 41 publications

Pixel-Wise Fabric Defect Detection by CNNs Without Labeled Training Data

Pixel-Wise Fabric Defect Detection by CNNs Without Labeled Training Data

The Algorithms of Image Processing and Analysis in the Textile Fabrics Abrasion Assessment

A One-Stage Approach for Surface Anomaly Detection with Background Suppression Strategies

Contact Info

Product

Resources

About