Improving Few-Shot Image Classification with Self-supervised Learning.

Shisheng Deng,Dongping Liao,Xitong Gao,Juanjuan Zhao,Kejiang Ye
DOI: https://doi.org/10.1007/978-3-031-23498-9_5
2022-01-01
Abstract:Few-Shot Image Classification (FSIC) aims to learn an image classifier with only a few training samples. The key challenge of few-shot image classification is to learn this classifier with scarce labeled data. To tackle the issue, we leverage the self-supervised learning (SSL) paradigm to exploit unsupervised information. This work builds upon two-stage training paradigm, to push the current state-of-the-art (SOTA) in solving FSIC problem further. Specifically, we incorporate the traditional self-supervised learning method (TSSL) into the pre-training stage and propose an episodic contrastive loss (CL) as an auxiliary supervision for the meta-training stage. The proposed bipartite method, called FSIC-SSL, can SOTA task accuracies on two mainstream FSIC benchmark datasets. Our code will be available at https://github.com/SethDeng/FSIC_SSL .
What problem does this paper attempt to address?