Few-shot Metric Learning: Online Adaptation of Embedding for Retrieval

TL;DR

This paper introduces a novel few-shot metric learning approach that enables online adaptation of embedding functions for improved image retrieval across diverse and unseen domains, especially when data is scarce.

Contribution

The paper proposes Channel-Rectifier Meta-Learning (CRML), a new method for online adaptation of metric spaces in few-shot scenarios, outperforming existing baselines on multiple datasets.

Findings

CRML improves metric learning performance in few-shot settings.

The method achieves significant gains when domain gaps are large.

Experimental results on multiple datasets validate the effectiveness of CRML.

Abstract

Metric learning aims to build a distance metric typically by learning an effective embedding function that maps similar objects into nearby points in its embedding space. Despite recent advances in deep metric learning, it remains challenging for the learned metric to generalize to unseen classes with a substantial domain gap. To tackle the issue, we explore a new problem of few-shot metric learning that aims to adapt the embedding function to the target domain with only a few annotated data. We introduce three few-shot metric learning baselines and propose the Channel-Rectifier Meta-Learning (CRML), which effectively adapts the metric space online by adjusting channels of intermediate layers. Experimental analyses on miniImageNet, CUB-200-2011, MPII, as well as a new dataset, miniDeepFashion, demonstrate that our method consistently improves the learned metric by adapting it to target classes and achieves a greater gain in image retrieval when the domain gap from the source classes is larger.