Semi-supervised defect classification of steel surface based on multi-training and generative adversarial network

He, Yu; Song, Kun; Dong, Hongwen; Yan, Yunhui

doi:10.1016/j.optlaseng.2019.06.020

Cited by 137 publications

(64 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Natarajan et al [31] proposed a flexible multi-layered deep feature extraction through transfer learning and SVM classifiers, which overcome the problem of over-fitting caused by small datasets. He et al [32] proposed a semi-supervised model of CNN for feature extraction and fed the representation features into a classifier for classification of steel surface defect. However, these methods can't give the exact location of defects.…”

Section: B Deep-learning-based Detection Approachesmentioning

confidence: 99%

PGA-Net: Pyramid Feature Fusion and Global Context Attention Network for Automated Surface Defect Detection

Dong

Song

et al. 2020

IEEE Trans. Ind. Inf.

Self Cite

349

View full text Add to dashboard Cite

Surface defect detection is a critical task in industrial production process. Nowadays, there are lots of detection methods based on computer vision and have been successfully applied in industry, they also achieved good results. However, achieving full automation of surface defect detection remains a challenge, due to the complexity of surface defect, in intra-class, while the defects between inter-class contain similar parts, there are large differences in appearance of the defects. To address these issues, this paper proposes a pyramid feature fusion and global context attention network for pixel-wise detection of surface defect, called PGA-Net. In the framework, the multi-scale features are extracted at first from backbone network. Then the pyramid feature fusion module is used to fuse these features into five resolutions through some efficient dense skip connections. Finally, the global context attention module is applied to the fusion feature maps of adjacent resolution, which allows effective information propagate from low-resolution fusion feature maps to high-resolution fusion ones. In addition, the boundary refinement block is added to the framework to refine the boundary of defect and improve the result of predict. The final prediction is the fusion of the five resolutions fusion feature maps. The results of evaluation on four real-world defect datasets demonstrate that the proposed method outperforms the state-of-the-art methods on mean Intersection of Union and mean Pixel Accuracy

show abstract

Section: B Deep-learning-based Detection Approachesmentioning

confidence: 99%

PGA-Net: Pyramid Feature Fusion and Global Context Attention Network for Automated Surface Defect Detection

Dong

Song

et al. 2020

IEEE Trans. Ind. Inf.

Self Cite

349

View full text Add to dashboard Cite

show abstract

“…Industrial applications gradually adopt deep learning methods to detect industrial products. Due to the smaller number of defect datasets on hot-rolled steel surfaces, some papers used semi-supervised methods to augment datasets [32]- [34]. For example, Yiping Gao et al raised a method, which used the semi-supervised framework to generate fake labels to satisfied the label's requirement.…”

Section: Related Theory Analysismentioning

confidence: 99%

“…Yu He et al showed that a categorized generative adversarial network (GAN) is used to generate a large amount of unmarked data against the network, and then the residual network is used for classification. The final classification accuracy rate is 99.56% [34]. In order to improve the detection accuracy, He, D et al used combined networks to detect surface defects, in which multi-group convolutional neural network (MG-CNN) was used to preclassification; and then different types of extracted defect features were input into another Yolo-based neural network for detection and recognition.…”

Section: Related Theory Analysismentioning

confidence: 99%

A Steel Surface Defect Recognition Algorithm Based on Improved Deep Learning Network Model Using Feature Visualization and Quality Evaluation

2020

View full text Add to dashboard Cite

Steel defect detection is used to detect defects on the surface of the steel and to improve the quality of the steel surface. However, traditional image detection algorithms cannot meet the detection requirements because of small defect features and low contrast between background and features about steel surface defect datasets. A novel recognition algorithm for steel surface defects based on improved deep learning network models using feature visualization and quality evaluation is proposed in this paper. Firstly, the VGG19 is used to pre-train the steel surface defect classification task and the corresponding DVGG19 is established to extract the feature images in different layers from defects weight model. Secondly, the SSIM and decision tree are used to evaluate the feature image quality and adjust the parameters and structure of VGG19. On this basis, a new VSD network is obtained and used for the classification of steel surface defects. Comparing with ResNet and VGG19 methods, experiment results show that the proposed method markedly can improve the average accuracy of classification, and the model is able to converge quickly, which can be good for steel surface defect recognition using VSD network model of feature visualization and quality evaluation.

show abstract

“…The images are obtained from the Northeastern University (NEU) surface database [20], [21], [22] which contains six types of defects (rolled-in scale (Rs), patches (Pa), crazing (Cr), pitted surface (Ps), inclusion (In) and scratches (Sc)) with 300 images for each defect (1800 total). Image size is 200×200 pixels with a .bmp format and the images are in grey-scale.…”

Section: Datasetmentioning

confidence: 99%

Surface Defect Detection Using YOLO Network

Hatab

Malekmohamadi

Amira

2020

Advances in Intelligent Systems and Computing

View full text Add to dashboard Cite

Detecting defects on surfaces such as steel, can be a challenging task because defects have complex and unique features. These defects happen in many production lines and differ between each one of these production lines. In order to detect these defects, the You Only Look Once (YOLO) detector which uses a Convolutional Neural Network (CNN), is used and received only minor modifications. YOLO is trained and tested on a dataset containing six kinds of defects to achieve accurate detection and classification. The network can also obtain the coordinates of the detected bounding boxes, giving the size and location of the detected defects. Since manual defect detection is expensive, labor-intensive and inefficient, this paper contributes to the sophistication and improvement of manufacturing processes. This system can be installed on chipsets and deployed to a factory line to greatly improve quality control and be part of smart internet of things (IoT) based factories in the future. YOLO achieves a respectable 70.66% mean average precision (mAP) despite the small dataset and minor modifications to the network.

show abstract

Semi-supervised defect classification of steel surface based on multi-training and generative adversarial network

Cited by 137 publications

References 13 publications

PGA-Net: Pyramid Feature Fusion and Global Context Attention Network for Automated Surface Defect Detection

PGA-Net: Pyramid Feature Fusion and Global Context Attention Network for Automated Surface Defect Detection

A Steel Surface Defect Recognition Algorithm Based on Improved Deep Learning Network Model Using Feature Visualization and Quality Evaluation

Surface Defect Detection Using YOLO Network

Contact Info

Product

Resources

About