Improving Scarce RS Data Classification With Independent Noise and Feature Mutual Exclusion

Binghao Liu, Yidi Bai

Abstract: Data shortage and real-time demand are two essential characteristics of remote sensing (RS) signal processing tasks. Existing works rarely focus on the scenario where all categories data are relatively scarce, which may be necessary in some occasions. To further explore the potential of lightweight deep neural network for scarce data RS image classification, we attempt to improve the model performance in such task with tiny proportion of the training data. Due to the information insufficiency caused by data shortage, we propose a statistically independent Gaussian noise-based feature augmentation (IGNA) module. A variational autoencoder is used to provide noise vectors which come from the same input space with feature vectors for classification, but independent of them. To alleviate the interclass feature confusion caused by the feature augmentation and enhance decision boundaries, we design a $3\hat {\Sigma } ^{1/2}$ area-based interclass mutual exclusion (ICME) strategy to enlarge distances between samples of different classes in feature space with contrastive loss. Extensive experiments are conducted to prove that our method can significantly improve the performance of lightweight models on scarce data RS classification task.