Rectifying the Shortcut Learning of Background for Few-Shot Learning

TL;DR

This paper identifies background as a shortcut in Few-Shot Learning that aids in-class classification but hampers generalization, and proposes a framework to extract foreground objects without extra supervision to improve FSL performance.

Contribution

The paper introduces COSOC, a novel framework that extracts foreground objects in images without extra supervision to address background shortcut learning in FSL.

Findings

COSOC effectively reduces background shortcut bias.

Experiments show improved generalization in FSL tasks.

Foreground extraction enhances classification accuracy.

Abstract

The category gap between training and evaluation has been characterised as one of the main obstacles to the success of Few-Shot Learning (FSL). In this paper, we for the first time empirically identify image background, common in realistic images, as a shortcut knowledge helpful for in-class classification but ungeneralizable beyond training categories in FSL. A novel framework, COSOC, is designed to tackle this problem by extracting foreground objects in images at both training and evaluation without any extra supervision. Extensive experiments carried on inductive FSL tasks demonstrate the effectiveness of our approaches.