arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2508.02476cs.CR
收藏

PoseGuard:带安全护栏的姿态引导生成

PoseGuard: Pose-Guided Generation with Safety Guardrails

Kongxin Wang, Jie Zhang, Peigui Qi, Kunsheng Tang, Tianwei Zhang, Wenbo Zhou

AI总结:

针对姿态引导视频生成存在的身份冒用、NSFW内容生成等安全风险,研究提出PoseGuard安全对齐框架,通过双目标训练与LoRA微调实现恶意姿态下的生成质量降级,可灵活适配新威胁,且能推广至面部关键点引导生成场景。

摘要:

姿态引导视频生成已成为创意产业的强大工具,Animate Anyone等框架就是例证。然而,基于特定姿态的条件生成会带来严重风险,例如身份冒用、隐私侵犯以及NSFW(不适宜工作场所)内容生成。为应对这些挑战,我们提出了**PoseGuard**,一种面向姿态引导生成的安全对齐框架。PoseGuard的设计思路是:在遇到恶意姿态时通过降低输出质量来抑制不安全生成,同时对良性输入保持高保真输出。我们将不安全姿态划分为三种典型类型:下跪或冒犯性敬礼等歧视性手势、会生成NSFW内容的性暗示姿态,以及模仿受版权保护的名人动作的姿态。PoseGuard采用结合生成保真度与安全对齐的双目标训练策略,并使用基于LoRA(低秩适配)的微调实现高效、轻参数量的更新。为确保能够适应不断演变的威胁,PoseGuard支持特定姿态的LoRA融合,在发现新的不安全姿态时可实现灵活的模块化更新。我们进一步证明了PoseGuard可推广至面部关键点引导生成场景。大量实验验证表明,PoseGuard能有效阻断不安全生成,同时保持良性输入的生成质量,且对微小的姿态变化具备鲁棒性。

英文摘要:

Pose-guided video generation has become a powerful tool in creative industries, exemplified by frameworks like Animate Anyone. However, conditioning generation on specific poses introduces serious risks, such as impersonation, privacy violations, and NSFW content creation. To address these challenges, we propose $\textbf{PoseGuard}$, a safety alignment framework for pose-guided generation. PoseGuard is designed to suppress unsafe generations by degrading output quality when encountering malicious poses, while maintaining high-fidelity outputs for benign inputs. We categorize unsafe poses into three representative types: discriminatory gestures such as kneeling or offensive salutes, sexually suggestive poses that lead to NSFW content, and poses imitating copyrighted celebrity movements. PoseGuard employs a dual-objective training strategy combining generation fidelity with safety alignment, and uses LoRA-based fine-tuning for efficient, parameter-light updates. To ensure adaptability to evolving threats, PoseGuard supports pose-specific LoRA fusion, enabling flexible and modular updates when new unsafe poses are identified. We further demonstrate the generalizability of PoseGuard to facial landmark-guided generation. Extensive experiments validate that PoseGuard effectively blocks unsafe generations, maintains generation quality for benign inputs, and remains robust against slight pose variations.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑