Exploring Cross-Domain Few-Shot Classification via Frequency-Aware Prompting

TL;DR

This paper introduces a Frequency-Aware Prompting method with mutual attention for Cross-Domain Few-Shot Learning, enhancing robustness by leveraging frequency cues and improving existing methods' performance.

Contribution

It proposes a novel frequency-aware prompting mechanism and mutual attention module that can be integrated into existing CD-FSL methods to improve robustness and accuracy.

Findings

Significant performance improvements on CD-FSL benchmarks.

Enhanced robustness by focusing on frequency cues.

Compatibility as a plug-and-play module for existing methods.

Abstract

Cross-Domain Few-Shot Learning has witnessed great stride with the development of meta-learning. However, most existing methods pay more attention to learning domain-adaptive inductive bias (meta-knowledge) through feature-wise manipulation or task diversity improvement while neglecting the phenomenon that deep networks tend to rely more on high-frequency cues to make the classification decision, which thus degenerates the robustness of learned inductive bias since high-frequency information is vulnerable and easy to be disturbed by noisy information. Hence in this paper, we make one of the first attempts to propose a Frequency-Aware Prompting method with mutual attention for Cross-Domain Few-Shot classification, which can let networks simulate the human visual perception of selecting different frequency cues when facing new recognition tasks. Specifically, a frequency-aware prompting mechanism is first proposed, in which high-frequency components of the decomposed source image are switched either with normal distribution sampling or zeroing to get frequency-aware augment samples. Then, a mutual attention module is designed to learn generalizable inductive bias under CD-FSL settings. More importantly, the proposed method is a plug-and-play module that can be directly applied to most off-the-shelf CD-FLS methods. Experimental results on CD-FSL benchmarks demonstrate the effectiveness of our proposed method as well as robustly improve the performance of existing CD-FLS methods. Resources at https://github.com/tinkez/FAP_CDFSC.