Convolutional Feature Aggregation Network With Self-Supervised Learning and Decision Fusion for SAR Target Recognition

Convolutional neural network (CNN) has been successfully employed for synthetic aperture radar automatic target recognition (SAR-ATR). Whereas, few labeled synthetic aperture radar (SAR) images cannot train a CNN model with strong generalization. In practice, the annotation of SAR images is often difficult and time-consuming, so we can usually collect few labeled and massive unlabeled SAR images. Here, we propose a convolutional feature aggregation network (CFANet) with self-supervised learning and decision fusion for SAR-ATR with few labeled and massive unlabeled data. The major contributions of CFANet are threefold. First, we develop to concatenate feature maps (FMs) of different convolutional layers to extract more discriminative feature. Second, the massive unlabeled SAR images with self-supervised pseudolabels are employed to pretrain CFANet, and then, the few labeled SAR images are used to fine-tune the model. By doing this, the information in both labeled and unlabeled SAR images are employed for the downstream ATR. Third, to effectively extract the information in different layers, the first-order and second-order statistical features of different layers are also used to train two extra classifiers. Then, we can obtain three pieces of soft classification results yielded by softmax layer of CFANet and two classifiers for a query SAR target image. These soft classification results are combined by weighted arithmetic average (WAA) rule whose weights are learned by minimizing the mean squared error (MSE) between fusion results and ground truth on labeled SAR target images. The developed CFANet model was tested on MSTAR and FuSARship datasets comprising about 5000 images. The experimental results demonstrate CFANet that can usually achieve the highest classification accuracy compared with a variety of related SAR-ATR methods.