The goal of few-shot image classification is to learn a classifier that can be well generalized to the unseen classes with a few available labeled samples. One major challenge for few-shot learning is ...

Hyperspectral imaging (HSI) effectively supports land cover (LC) classification in understanding nature–human interactions. The LC classification quality depends heavily on the input features.