1

Relative Position and Map Networks in Few-shot Learning for Image Classification

Few-shot learning is an important research topic in image classification, which aims to train robust classifiers to categorize images coming from new classes where only a few labeled samples are available. Recently, metric learning based methods have …