“…Jing Yang et al [31] conducted a study that examines state-of-the-art deep learning methods in defect detection. The study classied defects in various products, such as electronic components, pipes, 1,2,3,4,5,6,7,8,9,10 The authors are with the VNU University of Engineering and Technology, Vietnam, Email: anh.ph@vnu.edu.vn, duongtanrb@gmail.com, nguyenmaicg03@gmail.com, 147anhnt@gmail.com, thuhangkt01@gmail.com, thangluu3333@gmail.com, hieuvandang3210@gmail.com, dongtran.robotics@gmail.com, vanntt@vnu.edu.vn and tungbt@vnu.edu.vn 10 Corresponding author: tungbt@vnu.edu.vn welded parts, and textile materials, into categories and reviewed the main techniques and deep learning methods for defects, highlighting their characteristics, strengths, and shortcomings. Tian Wang et al [32] proposed a deep convolutional neural network (CNN) for automated quality visual inspection to control product quality for increased efficiency.…”