SRasP: Self-Reorientation Adversarial Style Perturbation for Cross-Domain Few-Shot Learning

TL;DR

This paper introduces SRasP, a novel style perturbation method for cross-domain few-shot learning that enhances model robustness and transferability by stabilizing style perturbations and encouraging flatter minima.

Contribution

The paper proposes a crop-global style perturbation network with a multi-objective optimization for improved domain transfer in few-shot learning.

Findings

Consistent performance improvements on multiple benchmarks

Enhanced model generalization to unseen domains

Stabilized training leading to flatter minima

Abstract

Cross-Domain Few-Shot Learning (CD-FSL) aims to transfer knowledge from a seen source domain to unseen target domains, serving as a key benchmark for evaluating the robustness and transferability of models. Existing style-based perturbation methods mitigate domain shift but often suffer from gradient instability and convergence to sharp minima.To address these limitations, we propose a novel crop-global style perturbation network, termed Self-Reorientation Adversarial \underline{S}tyle \underline{P}erturbation (SRasP). Specifically, SRasP leverages global semantic guidance to identify incoherent crops, followed by reorienting and aggregating the style gradients of these crops with the global style gradients within one image. Furthermore, we propose a novel multi-objective optimization function to maximize visual discrepancy while enforcing semantic consistency among global, crop, and adversarial features. Applying the stabilized perturbations during training encourages convergence toward flatter and more transferable solutions, improving generalization to unseen domains. Extensive experiments are conducted on multiple CD-FSL benchmarks, demonstrating consistent improvements over state-of-the-art methods.