Data augmentation by a CycleGAN-based extra-supervised model for nondestructive testing. (31st January 2022)

• Record Type: Journal Article
• Authors: Jiangsha, Ai; Tian, Lulu; Bai, Libing; Zhang, Jie
• Is Part Of: Measurement science & technology. Volume 33:Number 4(2022)
• Publication Date: 2022-01-31
• Subjects: CycleGANs -- data augmentation -- defects recognition -- non-destructive testing -- defects generation; Physical measurements -- Periodicals; Scientific apparatus and instruments -- Periodicals; Equipment and Supplies -- Periodicals; Science -- instrumentation -- Periodicals; Technology -- instrumentation -- Periodicals; Mesures physiques -- Périodiques; Physical measurements; Scientific apparatus and instruments; Periodicals; 502.87
• DOI: 10.1088/1361-6501/ac3ec3
• Languages: English
• ISSNs: 0957-0233

Abstract

Abstract: The deep learning method is widely used in computer vision tasks with large-scale annotated datasets. However, obtaining such datasets in most directions of the vision based nondestructive testing (NDT) field is very challenging. Data augmentation is proved as an efficient way of dealing with the lack of large-scale annotated datasets. In this paper, we propose a CycleGAN-based extra-supervised (CycleGAN-ES) model to generate synthetic NDT images, where the ES is used to ensure that the bidirectional mapping is learned for corresponding labels and defects. Furthermore, we show the effectiveness of using the synthesized images to train deep convolutional neural networks (DCNNs) for defect recognition. In the experiments, we extract a number of x-ray welding images with both defect and no defects from the published GDXray dataset, and CycleGAN-ES is used to generate the synthetic defect images based on a small number of extracted defect images and manually drawn labels that are used as a content guide. For quality verification of the synthesized defect images, we use a high-performance classifier pretrained using a big dataset to recognize the synthetic defects and show the comparability of the performances of classifiers trained using synthetic defects and real defects, respectively. To present the effectiveness of using the synthesized defects as an augmentation method, we train and evaluate the performances of DCNN for defect recognition with or without the synthesized defects.