Class Shared Dictionary Learning for Few-Shot Remote Sensing Scene Classification

In the field of remote sensing, it is infeasible to collect a large number of labeled samples due to imaging equipment and imaging environment. Few-shot learning (FSL) is the dominant method to alleviate this problem, which pursues quickly adapting to novel categories from a limited number of labeled samples. The few-shot remote sensing scene classification (RSSC) generally includes the pretraining and meta-test phases. However, a “negative transfer” problem exists that data categories in both the phases are different. It causes the pretrained feature extractor to be unable well-adapted to the novel data category. This letter proposes class shared dictionary learning (CSDL) for few-shot RSSC to address this issue. Specifically, this letter designs the mirror-based feature extractor (MFE) in the pretraining phase, constructing a self-supervised classification task to improve the feature extractor robustness. Furthermore, this letter proposes a class shared dictionary (CSD) classifier based on dictionary learning. The CSD projects the novel data feature in meta-test into subspace to reconstruct more discriminative features and complete the classification task. Extensive experiments on remote sensing datasets have demonstrated that the proposed CSDL achieves advanced classification performance.