Data Augmentation using Style Transfer in SAR Automatic Target Classification

In this paper, a method is proposed which uses style transfer techniques in the field of computer vision to combine a small number of millimeter-wave images with a large number of optical images to generate a library of pseudo millimeter-wave images. Specifically, the style transfer method combines the style features of a millimeter-wave image with the content features of an optical image to generate a new image. By combining different style images and content images, a large number of new images can be generated. The above generated images are then used to train any deep network for classification. The performance of proposed method is compared with a conventional method of data augmentation. In addition, we testified the conventional transfer learning (TL) based data augmentation as well. The comparison results show that the method proposed in this paper effectively improves the accuracy of automatic classification in SAR automatic target classification.