I. Introduction
Quality inspection is an important part of industrial manufacturing [1], [2]. For industrial products such as steel, printed circuit boards (PCBs), and ceramics, surface defects generated during production not only affect appearance, but also have security risks [3]. Thus, detecting surface defect is essential to industrial production [4]. In recent years, with the development of Internet of Things (IoT) technology, deep learning methods have been widely used in smart manufacturing, and computer vision-based automated surface defect detection is the trend of future development.