English

TAIGen: Training-Free Adversarial Image Generation via Diffusion Models

Computer Vision and Pattern Recognition 2025-08-22 v1 Artificial Intelligence

Abstract

Adversarial attacks from generative models often produce low-quality images and require substantial computational resources. Diffusion models, though capable of high-quality generation, typically need hundreds of sampling steps for adversarial generation. This paper introduces TAIGen, a training-free black-box method for efficient adversarial image generation. TAIGen produces adversarial examples using only 3-20 sampling steps from unconditional diffusion models. Our key finding is that perturbations injected during the mixing step interval achieve comparable attack effectiveness without processing all timesteps. We develop a selective RGB channel strategy that applies attention maps to the red channel while using GradCAM-guided perturbations on green and blue channels. This design preserves image structure while maximizing misclassification in target models. TAIGen maintains visual quality with PSNR above 30 dB across all tested datasets. On ImageNet with VGGNet as source, TAIGen achieves 70.6% success against ResNet, 80.8% against MNASNet, and 97.8% against ShuffleNet. The method generates adversarial examples 10x faster than existing diffusion-based attacks. Our method achieves the lowest robust accuracy, indicating it is the most impactful attack as the defense mechanism is least successful in purifying the images generated by TAIGen.

Keywords

Cite

@article{arxiv.2508.15020,
  title  = {TAIGen: Training-Free Adversarial Image Generation via Diffusion Models},
  author = {Susim Roy and Anubhooti Jain and Mayank Vatsa and Richa Singh},
  journal= {arXiv preprint arXiv:2508.15020},
  year   = {2025}
}

Comments

Accepted at ICCVW-CV4BIOM 2025

R2 v1 2026-07-01T04:59:02.606Z