A Closer Look at Few-shot Classification

Abstract

Few-shot classi?cation aims to learn a classi?er to recognize unseen classes during training with limited labeled examples. While signi?cant progress has been made, the growing complexity of network designs, meta-learning algorithms, and differences in implementation details make a fair comparison dif?cult. In this paper, we present 1) a consistent comparative analysis of several representative few-shot classi?cation algorithms, with results showing that deeper backbones signi?cantly reduce the gap across methods including the baseline, 2) a slightly modi?ed baseline method that surprisingly achieves competitive performance when compared with the state-of-the-art on both the mini-ImageNet and the CUB datasets, and 3) a new experimental setting for evaluating the cross-domain generalization ability for few-shot classi?cation algorithms. Our results reveal that reducing intra-class variation is an important factor when the feature backbone is shallow, but not as critical when using deeper backbones. In a realistic, cross-domain evaluation setting, we show that a baseline method with a standard ?ne-tuning practice compares favorably against other state-of-the-art few-shot learning algorithms.