Towards Effective and Efficient Adversarial Defense with Diffusion Models for Robust Visual Tracking

Abstract

Although deep learning-based visual tracking methods have made significant progress, they exhibit vulnerabilities when facing carefully designed adversarial attacks, which can lead to a sharp decline in tracking performance. To address this issue, this paper proposes for the first time a novel adversarial defense method based on denoise diffusion probabilistic models, termed DiffDf, aimed at effectively improving the robustness of existing visual tracking methods against adversarial attacks. DiffDf establishes a multi-scale defense mechanism by combining pixel-level reconstruction loss, semantic consistency loss, and structural similarity loss, effectively suppressing adversarial perturbations through a gradual denoising process. Extensive experimental results on several mainstream datasets show that the DiffDf method demonstrates excellent generalization performance for trackers with different architectures, significantly improving various evaluation metrics while achieving real-time inference speeds of over 30 FPS, showcasing outstanding defense performance and efficiency. Codes are available at https://github.com/pgao-lab/DiffDf.

References in corpus (11)

• Explaining and Harnessing Adversarial Examples
• Denoising Diffusion Probabilistic Models
• Intriguing properties of neural networks
• Feature Squeezing: Detecting Adversarial Examples in Deep Neural Networks
• GOT-10k: A Large High-Diversity Benchmark for Generic Object Tracking in the Wild
• ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models
• Adversarially Robust Distillation
• Diffusion Models for Adversarial Purification
• Stimulating Diffusion Model for Image Denoising via Adaptive Embedding and Ensembling
• Guided Diffusion Model for Adversarial Purification
• Improving Adversarial Robustness via Attention and Adversarial Logit Pairing

Cited by in corpus (1)

• ASBI: Leveraging Informative Real-World Data for Active Black-Box Simulator Tuning