A Weakly Supervised Surface Defect Detection Based on Convolutional Neural Network

Abstract

Surface defect detection is a critical task in product quality assurance for manufacturing lines. The deep learning-based methods recently developed for defect detection are typically trained using a supervised learning strategy and large defect sample sets. Conventional methods often require additional pixel-level labeling or bounding boxes to predict the location of defects. However, the number of required samples and the time-intensive annotation process limits the practical use of these algorithms. As such, this study proposes a weakly supervised detection framework in which a CNN model is trained to identify surface cracks in motor commutators. The model was trained using small subsets of defect samples (∼5-30) and does not require a pre-trained network. This approach consists of localization and decision networks that simultaneously predict both the location and probability of defects. A new loss function was also developed to identify abnormal regions in a sample with accessible image-level labels. A collaboration learning strategy was then applied to utilize the loss function and compensate for imbalances at the pixel level. Experimental results using a small number of image-level training labels from a real industrial dataset exhibited a 99.5% recognition accuracy, which is comparable to relevant methods using pixel-level labels.

A Review and Analysis of Automatic Optical Inspection and Quality Monitoring Methods in Electronics Industry

“…Deep learning has recently become the most influential technology in machine vision and pattern recognition problems [383]. Deep learning concept involves the usage of Deep Neural Networks (DNN), which can handle feature extraction and classification methods in machine vision problems.…”

“…Remarkable achievements in feature extraction and image classification have been produced by CNNs such as AlexNet [394], VGG [395], ResNet [396], and DenseNet [397], which have outperformed conventional classification models such as SVM. As a result, CNN was applied to optical defect detection problems after AlexNet was proposed [383]. A typical CNN consists of input layer, convolutional layers, pooling layers, fully connected layers, and output layer as shown in Figure 44.…”

Deep learning for industrial image: challenges, methods for enriching the sample space and restricting the hypothesis space, and possible issue

“…In the work by Xu et al (2020), 3-fold cross validation is used to ensure that the same images do not appear in both the training and test sets. 4-fold cross validation is used in reference (Y. X.…”

“…Therefore, the optimization space can be smoother, thereby improving the generalization ability of the DLMs. This method has been widely used in the industrial field (Akram et al 2020; Gonzalez-Val et al 2020;Xu et al 2020;Z. P. Guo et al 2019;Q.…”

Surface Defects Inspection of Cylindrical Metal Workpieces Based on Weakly Supervised Learning

“…Chen [24] et al proposed a robust weakly supervised learning method for surface defect detection, which uses transfer learning to get CAM. Xu [25] et al proposed a weakly supervised detection framework, in which CNN model was trained to identify surface cracks in motor commutators. The method achieved 99.5% recognition accuracy in Kolektor SSD dataset.…”