Recognition Of Surface Defects On Steel Sheet Using Transfer Learning
This work addresses automatic defect recognition in steel production, offering a solution for scenarios with few data, but it is incremental as it builds on existing transfer learning and CNN methods.
The paper tackled the problem of recognizing surface defects on steel sheets with limited training data by proposing a transfer learning approach using part of pretrained VGG16 as a feature extractor and a new CNN as a classifier, achieving accuracies of 99.1% and 96.0% on datasets with 150 and 10 images per class, respectively.
Automatic defect recognition is one of the research hotspots in steel production, but most of the current methods mainly extract features manually and use machine learning classifiers to recognize defects, which cannot tackle the situation, where there are few data available to train and confine to a certain scene. Therefore, in this paper, a new approach is proposed which consists of part of pretrained VGG16 as a feature extractor and a new CNN neural network as a classifier to recognize the defect of steel strip surface based on the feature maps created by the feature extractor. Our method achieves an accuracy of 99.1% and 96.0% while the dataset contains 150 images each class and 10 images each class respectively, which is much better than previous methods.