[2603.05135] SRasP: Self-Reorientation Adversarial Style Perturbation for Cross-Domain Few-Shot Learning

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.05135 (cs) [Submitted on 5 Mar 2026] Title:SRasP: Self-Reorientation Adversarial Style Perturbation for Cross-Domain Few-Shot Learning Authors:Wenqian Li, Pengfei Fang, Hui Xue View a PDF of the paper titled SRasP: Self-Reorientation Adversarial Style Perturbation for Cross-Domain Few-Shot Learning, by Wenqian Li and 2 other authors View PDF HTML (experimental) 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 this http URL 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, improv...