“…In recent years, deep learning-based methods such as the convolutional neural network (CNN) have rapidly replaced conventional pattern recognition methods as they dramatically improved the performances by automatically learning the discriminative features for SAR target recognition. In most relevant studies, excellent SAR target recognition performances were achieved by the combination of the CNN and pattern recognition [ 10 , 11 , 12 , 13 ], extraction of multi-view features [ 14 , 15 , 16 ], adoption of state-of-the-art deep learning network structures [ 17 , 18 , 19 , 20 , 21 , 22 ], fusion of feature maps [ 23 ], data augmentation [ 24 , 25 ], and transfer learning [ 26 , 27 ].…”