FSSDD: Few‐shot steel defect detection based on multi‐scale semantic enhancement representation and mask category information mapping

Steel defect detection is important for industry production as it is tied to the product quality and production efficiency. However, previous steel defect detection methods based on deep convolutional neural networks heavily rely on large‐scale data for training and tend to have poor generalization ability for a novel defect category. In this paper, a novel few‐shot steel defect detection model based on multi‐scale semantic enhancement representation and mask category information mapping is introduced, where only a few annotated samples are acquired for the novel defect category. More concretely, three main components are built: an information‐guidance enhanced multi‐head detector is proposed to improve the representation of information in meta‐feature maps, a mask category representation module is designed to enhance the category feature representation of the mask region in the support set, and a novel multi‐scale category edge loss function is designed to assist the generation of category reweighting vector. Extensive experiments on the North‐east University few‐shot steel defect data set demonstrate that the proposed method significantly outperforms the state‐of‐the‐art methods and verify its effectiveness through ablation studies. 1.In this paper, we introduce a few‐shot defect detection method based on meta‐learning framework on steel surface images (FSSDD). Our method learns transferable meta‐features from large amount of base classes and extends them by fine‐tuning to new classes with only a few labeled samples. 2.We propose an information‐guided enhanced multi‐head detector (IGEM) to enhance the information expression of meta‐features on query images. Meanwhile, a mask category representation module (MCR) and a multi‐scale class edge (MSCE) loss function are designed to improve the representation ability of meta‐learning framework on each category. 3.Extensive experiments illustrate that our approach has achieved consistent improvement on few‐shot steel defect data set. Specially, our approach achieves better performance than the state‐of‐the‐art methods on the data set.