SyncBreaker:Stage-Aware Multimodal Adversarial Attacks on Audio-Driven Talking Head Generation
2026-04-09 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors created SyncBreaker, a tool to stop fake talking-head videos made from both audio and images from working properly. Unlike past methods that only protect either the image or the audio, SyncBreaker tweaks both inputs to confuse the system that animates faces with speech. They use special techniques to keep the image looking normal while messing up how the audio controls lip movements. Tests show SyncBreaker disrupts lip-sync better than other methods without making the original inputs look bad.
diffusion-based generationtalking-head animationlip synchronizationmulti-modal protectionaudio-driven animationcross-attentiondenoising diffusionperceptual constraintswhite-box settingface forgery detection
Authors
Wenli Zhang, Xianglong Shi, Sirui Zhao, Xinqi Chen, Guo Cheng, Yifan Xu, Tong Xu, Yong Liao
Abstract
Diffusion-based audio-driven talking-head generation enables realistic portrait animation, but also introduces risks of misuse, such as fraud and misinformation. Existing protection methods are largely limited to a single modality, and neither image-only nor audio-only attacks can effectively suppress speech-driven facial dynamics. To address this gap, we propose SyncBreaker, a stage-aware multimodal protection framework that jointly perturbs portrait and audio inputs under modality-specific perceptual constraints. Our key contributions are twofold. First, for the image stream, we introduce nullifying supervision with Multi-Interval Sampling (MIS) across diffusion stages to steer the generation toward the static reference portrait by aggregating guidance from multiple denoising intervals. Second, for the audio stream, we propose Cross-Attention Fooling (CAF), which suppresses interval-specific audio-conditioned cross-attention responses. Both streams are optimized independently and combined at inference time to enable flexible deployment. We evaluate SyncBreaker in a white-box proactive protection setting. Extensive experiments demonstrate that SyncBreaker more effectively degrades lip synchronization and facial dynamics than strong single-modality baselines, while preserving input perceptual quality and remaining robust under purification. Code: https://github.com/kitty384/SyncBreaker.