Generating Transferrable Adversarial Examples via Local Mixing and Logits Optimization for Remote Sensing Object Recognition

2025-09-09

AI Summary

In remote sensing image recognition, deep neural networks are vulnerable to adversarial attacks; existing hybrid strategies impair semantic consistency either through global aliasing or region-wise hard swapping, while cross-entropy loss induces gradient vanishing during iterative optimization, limiting transferability. To address these issues, this paper proposes a synergistic adversarial generation framework integrating local-region-based mixing and logits-level optimization: (i) a localized MixUp mechanism preserves global semantic coherence; (ii) logits loss replaces cross-entropy to mitigate gradient decay; and (iii) a smoothing loss suppresses high-frequency perturbation noise. Evaluated on FGSCR-42 and MTARSI benchmarks, the method significantly enhances cross-model transferability—achieving an average 17.28% improvement in black-box attack success rate and outperforming 12 state-of-the-art methods.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Adversarial Learning & Robustness
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Deep Neural Networks (DNNs) are vulnerable to adversarial attacks, posing significant security threats to their deployment in remote sensing applications. Research on adversarial attacks not only reveals model vulnerabilities but also provides critical insights for enhancing robustness. Although current mixing-based strategies have been proposed to increase the transferability of adversarial examples, they either perform global blending or directly exchange a region in the images, which may destroy global semantic features and mislead the optimization of adversarial examples. Furthermore, their reliance on cross-entropy loss for perturbation optimization leads to gradient diminishing during iterative updates, compromising adversarial example quality. To address these limitations, we focus on non-targeted attacks and propose a novel framework via local mixing and logits optimization. First, we present a local mixing strategy to generate diverse yet semantically consistent inputs. Different from MixUp, which globally blends two images, and MixCut, which stitches images together, our method merely blends local regions to preserve global semantic information. Second, we adapt the logit loss from targeted attacks to non-targeted scenarios, mitigating the gradient vanishing problem of cross-entropy loss. Third, a perturbation smoothing loss is applied to suppress high-frequency noise and enhance transferability. Extensive experiments on FGSCR-42 and MTARSI datasets demonstrate superior performance over 12 state-of-the-art methods across 6 surrogate models. Notably, with ResNet as the surrogate on MTARSI, our method achieves a 17.28% average improvement in black-box attack success rate.

Problem

• Generating transferable adversarial examples for remote sensing object recognition
• Addressing gradient vanishing in cross-entropy loss during perturbation optimization
• Preserving global semantic features while creating diverse adversarial inputs

Innovation

• Local mixing strategy preserves global semantic features
• Logit loss adaptation prevents gradient vanishing issues
• Perturbation smoothing suppresses noise and boosts transferability