Improved Few-Shot Visual Classification

TL;DR

This paper introduces Simple CNAPS, a new few-shot learning method that uses a Mahalanobis distance metric and adaptive feature extractors, achieving significant performance gains with fewer parameters.

Contribution

It demonstrates that a simple class-covariance-based metric combined with learnable feature extractors can outperform more complex approaches in few-shot image classification.

Findings

Up to 6.1% performance improvement over state-of-the-art

Simple CNAPS has 9.2% fewer trainable parameters

Effective estimation of high-dimensional covariances from few samples

Abstract

Few-shot learning is a fundamental task in computer vision that carries the promise of alleviating the need for exhaustively labeled data. Most few-shot learning approaches to date have focused on progressively more complex neural feature extractors and classifier adaptation strategies, as well as the refinement of the task definition itself. In this paper, we explore the hypothesis that a simple class-covariance-based distance metric, namely the Mahalanobis distance, adopted into a state of the art few-shot learning approach (CNAPS) can, in and of itself, lead to a significant performance improvement. We also discover that it is possible to learn adaptive feature extractors that allow useful estimation of the high dimensional feature covariances required by this metric from surprisingly few samples. The result of our work is a new "Simple CNAPS" architecture which has up to 9.2% fewer trainable parameters than CNAPS and performs up to 6.1% better than state of the art on the standard few-shot image classification benchmark dataset.