A Unified Framework for Stealthy Adversarial Generation via Latent Optimization and Transferability Enhancement
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913919265669120 |
|---|---|
| author | Pei, Gaozheng Ma, Ke Zhang, Dongpeng Sun, Chengzhi Xu, Qianqian Huang, Qingming |
| author_facet | Pei, Gaozheng Ma, Ke Zhang, Dongpeng Sun, Chengzhi Xu, Qianqian Huang, Qingming |
| contents | Due to their powerful image generation capabilities, diffusion-based adversarial example generation methods through image editing are rapidly gaining popularity. However, due to reliance on the discriminative capability of the diffusion model, these diffusion-based methods often struggle to generalize beyond conventional image classification tasks, such as in Deepfake detection. Moreover, traditional strategies for enhancing adversarial example transferability are challenging to adapt to these methods. To address these challenges, we propose a unified framework that seamlessly incorporates traditional transferability enhancement strategies into diffusion model-based adversarial example generation via image editing, enabling their application across a wider range of downstream tasks. Our method won first place in the "1st Adversarial Attacks on Deepfake Detectors: A Challenge in the Era of AI-Generated Media" competition at ACM MM25, which validates the effectiveness of our approach. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_23676 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Unified Framework for Stealthy Adversarial Generation via Latent Optimization and Transferability Enhancement Pei, Gaozheng Ma, Ke Zhang, Dongpeng Sun, Chengzhi Xu, Qianqian Huang, Qingming Computer Vision and Pattern Recognition Due to their powerful image generation capabilities, diffusion-based adversarial example generation methods through image editing are rapidly gaining popularity. However, due to reliance on the discriminative capability of the diffusion model, these diffusion-based methods often struggle to generalize beyond conventional image classification tasks, such as in Deepfake detection. Moreover, traditional strategies for enhancing adversarial example transferability are challenging to adapt to these methods. To address these challenges, we propose a unified framework that seamlessly incorporates traditional transferability enhancement strategies into diffusion model-based adversarial example generation via image editing, enabling their application across a wider range of downstream tasks. Our method won first place in the "1st Adversarial Attacks on Deepfake Detectors: A Challenge in the Era of AI-Generated Media" competition at ACM MM25, which validates the effectiveness of our approach. |
| title | A Unified Framework for Stealthy Adversarial Generation via Latent Optimization and Transferability Enhancement |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.23676 |