Deep Learning for Target Classification from SAR Imagery: Data Augmentation and Translation Invariance

This report deals with translation invariance of convolutional neural (CNNs) for automatic target recognition (ATR) from synthetic aperture (SAR) imagery. In particular, the translation invariance of CNNs for SAR represents the robustness against misalignment of target chips extracted SAR images. To understand the translation invariance of the CNNs, we CNNs which classify the target chips from the MSTAR into the ten under the condition of with and without data augmentation, and then the translation invariance of the CNNs. According to our results, if we use a deep residual network, the translation invariance of the CNN data augmentation using the aligned images such as the MSTAR target is not so large. A more important factor of translation invariance is the of augmented training data. Furthermore, our CNN using augmented training achieved a state-of-the-art classification accuracy of 99.6%. These show an importance of domain-specific data augmentation.