Surface anomaly detection using feature-based transfer learning for IoT-enabled smart manufacturing

Owing to the advancement of computational technology, the employment of deep learning architecture for defect detection in the manufacturing industry has gained considerable attention. Traditional means of defect detection through manual vis-ual inspection by operators are laborious as well as susceptible to mistakes. In the present study, a feature-based transfer learning approach is used to classify sur-face defects. The KolektorSDD database is used in the present study. Two pipe-lines were developed to investigate their efficacy in detecting the defects, namely the InceptionV3-SVM and VGG19-SVM pipelines, respectively. It was demon-strated from the study that the VGG19-SVM pipeline could provide desirable re-sults compared to the InceptionV3-SVM pipeline, suggesting that the VGG19 is a better feature extractor for the evaluated surface defects. It could be concluded that the proposed pipeline is suitable for the classification of surface defects.