Efficient and Robust Video Defense Framework against 3D-field Personalized Talking Face

Rui-qing Sun1, Xingshan Yao1, Tian Lan2, Jia-Ling Shi1, Chen-Hao Cui1, Hui-Yang Zhao1, Zhijing Wu1, Chen Yang1,*, Xian-Ling Mao1,*
1Beijing Institute of Technology, 2Alibaba International Digital Commerce

Video Demonstration

Demonstration of our video defense framework. The video shows the effectiveness of our method in protecting portrait videos while maintaining visual quality.

Abstract

State-of-the-art 3D-field video-referenced Talking Face Generation (TFG) methods synthesize high-fidelity personalized talking-face videos in real time by modeling 3D geometry and appearance from reference portrait video. This capability raises significant privacy concerns regarding malicious misuse of personal portraits. However, no efficient defense framework exists to protect such videos against 3D-field TFG methods. While image-based defenses could apply per-frame 2D perturbations, they incur prohibitive computational costs, severe video quality degradation, failing to disrupt 3D information for video protection.

To address this, we propose a novel and efficient video defense framework against 3D-field TFG methods, which protects portrait video by perturbing the 3D information acquisition process while maintaining high-fidelity video quality. Specifically, our method introduces: (1) a similarity-guided parameter sharing mechanism for computational efficiency, and (2) a multi-scale dual-domain attention module to jointly optimize spatial-frequency perturbations. Extensive experiments demonstrate that our proposed framework exhibits strong defense capability and achieves a 47× acceleration over the fastest baseline while maintaining high fidelity. Moreover, it remains robust against scaling operations and state-of-the-art purification attacks, and the effectiveness of our design choices is further validated through ablation studies.

Framework Overview

VDF Framework Architecture
Overall architecture of the proposed framework. (a) Efficiency improvement mechanism. (b) Dual-domain attention mechanism with multiscale module.

BibTeX

@article{sun2024efficient,
  title={Efficient and Robust Video Defense Framework against 3D-field Personalized Talking Face},
  author={Sun, Rui-qing and Yao, Xingshan and Lan, Tian and Shi, Jia-Ling and Cui, Chen-Hao and Zhao, Hui-Yang and Wu, Zhijing and Yang, Chen and Mao, Xian-Ling},
  journal={arXiv preprint arXiv:2512.21019},
  year={2024}
}