A Light-Weight Deep-Learning Model with Multi-Scale Features for Steel Surface Defect Classification

Liu, Yang; Yuan, Yachao; Balta, Cristhian; Liu, Jing

doi:10.3390/ma13204629

Cited by 32 publications

(27 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Liu et al [ 39 ] constructed a light‐weight DL model and performed classification experiments on the NEU dataset, and achieved the classification accuracy rate of 98.89%. Wu et al [ 40 ] resorted the transfer learning model to make the steel strip surface inspection, and its accuracy of training and testing reached about 98%.…”

Section: Experiments and Discussionmentioning

confidence: 99%

“…Di et al [28] also invent a semisupervised learning network called convolutional autoencoder and semisupervised generative adversarial networks (CAE-SGAN) to classify the surface defects of steel strip, and try to improve the performance of the model through the limited training samples, but the final classification accuracy was only 96.5%. Liu et al [39] constructed a light-weight DL model and performed classification experiments on the NEU dataset, and achieved the classification accuracy rate of 98.89%. Wu et al [40] resorted the transfer learning model to make the steel strip surface inspection, and its accuracy of training and testing reached about 98%.…”

Section: Comparisons With State-of-the-artsmentioning

confidence: 99%

“…Ashour et al [38] 96.00 Unknown Zhou et al [45] 94.91 Unknown He et al [27] 99.56 Unknown Cohn et al [24] 99.40 Unknown Di et al [28] 96.50 Unknown Wan et al [26] 97.50 Unknown Liu et al [39] 98.89 5.58 Without pretraining 90.91 3.17…”

Section: Comparisons With State-of-the-artsmentioning

confidence: 99%

See 2 more Smart Citations

Surface Defect Classification of Steel Strip with Few Samples Based on Dual‐Stream Neural Network

Zhang

Yan

et al. 2021

steel research int.

View full text Add to dashboard Cite

The automatic classification of surface defects is significant to the steel strip inspection system. However, influenced by the lack of samples, it is difficult to improve the classification accuracy. Herein, a novel dual‐stream neural network is proposed, which is composed by two streams of sample generation and classification training. Subsequently, numerous defect samples are generated for the classifier pretraining, and the real steel strip surface defects are classified by the transfer learning method. The experiment results demonstrate that the dual‐stream neural network has the highest classification accuracy of 99.70% among the state‐of‐the‐art classifiers, and the classification accuracy increasesd by 9.67% through generating plentiful similar images for pretraining. Moreover, the convolutional neural network (CNN) can improve the quality of generated defect images. When the number of epochs reaches 2000, the generated images are closest to the real labeled images. Finally, the cross‐validation proves that the dual‐stream neural network has excellent stability and robustness, without over‐fitting problem in the training process.

show abstract

Section: Experiments and Discussionmentioning

confidence: 99%

Section: Comparisons With State-of-the-artsmentioning

confidence: 99%

See 1 more Smart Citation

Surface Defect Classification of Steel Strip with Few Samples Based on Dual‐Stream Neural Network

Zhang

Yan

et al. 2021

steel research int.

View full text Add to dashboard Cite

show abstract

“…In continuation of deep learning frameworks on defect recognition, an automatic surface defects recognition model for steel strip production quality control using an effective compact convolutional neural network [ 9 , 20 , 21 ] to learn the low-level features present in the steel strips with the aid of multiple receptive fields was proposed in a study by [ 20 , 21 ]. In another related work, [ 22 ] proposed a lightweight Concurrent Convolutional Neural Network (ConCNN) to learn and classify steel surface multi-scale features for rapid real-time defect visual inspection and quality control. Also, a welding X-ray image frame defect classification was conducted using varied deep learning networks’ parameters and hyper-parameters in conjunction with other image processing algorithms such as Canny edge detection and adaptive Gaussian threshold methods [ 23 ].…”

Section: Related Workmentioning

confidence: 99%

A Hard Voting Policy-Driven Deep Learning Architectural Ensemble Strategy for Industrial Products Defect Recognition and Classification

Stephen

Madanian

Nguyen

2022

Sensors

View full text Add to dashboard Cite

Manual or traditional industrial product inspection and defect-recognition models have some limitations, including process complexity, time-consuming, error-prone, and expensiveness. These issues negatively impact the quality control processes. Therefore, an efficient, rapid, and intelligent model is required to improve industrial products’ production fault recognition and classification for optimal visual inspections and quality control. However, intelligent models obtained with a tradeoff of high accuracy for high latency are tedious for real-time implementation and inferencing. This work proposes an ensemble deep-leaning architectural framework based on a deep learning model architectural voting policy to compute and learn the hierarchical and high-level features in industrial artefacts. The voting policy is formulated with respect to three crucial viable model characteristics: model optimality, efficiency, and performance accuracy. In the study, three publicly available industrial produce datasets were used for the proposed model’s various experiments and validation process, with remarkable results recorded, demonstrating a significant increase in fault recognition and classification performance in industrial products. In the study, three publicly available industrial produce datasets were used for the proposed model’s various experiments and validation process, with remarkable results recorded, demonstrating a significant increase in fault recognition and classification performance in industrial products.

show abstract

“…Recently, with the development of industrial artificial intelligence along with computer technology, many studies have been reported using artificial intelligence to detect surface defects. The convolutional neural network (CNN) is an image-based deep learning algorithm and is a representative model used to surface inspection [ 4 , 5 , 6 , 7 , 8 , 9 ]. By repetitive training, features that define surface defects are automatically extracted without expert assistance.…”

Section: Introductionmentioning

confidence: 99%

Analysis of the Region of Interest According to CNN Structure in Hierarchical Pattern Surface Inspection Using CAM

et al. 2021

View full text Add to dashboard Cite

A convolutional neural network (CNN), which exhibits excellent performance in solving image-based problem, has been widely applied to various industrial problems. In general, the CNN model was applied to defect inspection on the surface of raw materials or final products, and its accuracy also showed better performance compared to human inspection. However, surfaces with heterogeneous and complex backgrounds have difficulties in separating defects region from the background, which is a typical challenge in this field. In this study, the CNN model was applied to detect surface defects on a hierarchical patterned surface, one of the representative complex background surfaces. In order to optimize the CNN structure, the change in inspection performance was analyzed according to the number of layers and kernel size of the model using evaluation metrics. In addition, the change of the CNN’s decision criteria according to the change of the model structure was analyzed using a class activation map (CAM) technique, which can highlight the most important region recognized by the CNN in performing classification. As a result, we were able to accurately understand the classification manner of the CNN for the hierarchical pattern surface, and an accuracy of 93.7% was achieved using the optimized model.

show abstract

A Light-Weight Deep-Learning Model with Multi-Scale Features for Steel Surface Defect Classification

Cited by 32 publications

References 27 publications

Surface Defect Classification of Steel Strip with Few Samples Based on Dual‐Stream Neural Network

Surface Defect Classification of Steel Strip with Few Samples Based on Dual‐Stream Neural Network

A Hard Voting Policy-Driven Deep Learning Architectural Ensemble Strategy for Industrial Products Defect Recognition and Classification

Analysis of the Region of Interest According to CNN Structure in Hierarchical Pattern Surface Inspection Using CAM

Contact Info

Product

Resources

About