PoseGuard: Pose-Guided Generation with Safety Guardrails

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Kongxin, Zhang, Jie, Qi, Peigui, Tang, Kunsheng, Zhang, Tianwei, Zhou, Wenbo
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916879830876160
author Wang, Kongxin
Zhang, Jie
Qi, Peigui
Tang, Kunsheng
Zhang, Tianwei
Zhou, Wenbo
author_facet Wang, Kongxin
Zhang, Jie
Qi, Peigui
Tang, Kunsheng
Zhang, Tianwei
Zhou, Wenbo
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2508_02476
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PoseGuard: Pose-Guided Generation with Safety Guardrails
Wang, Kongxin
Zhang, Jie
Qi, Peigui
Tang, Kunsheng
Zhang, Tianwei
Zhou, Wenbo
Cryptography and Security
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.
title PoseGuard: Pose-Guided Generation with Safety Guardrails
topic Cryptography and Security
url https://arxiv.org/abs/2508.02476