# SyncBreaker:Stage-Aware Multimodal Adversarial Attacks on Audio-Driven Talking Head Generation

Source: [SyncBreaker:Stage-Aware Multimodal Adversarial Attacks on Audio-Driven Talking Head Generation](<https://www.sota2.com/research/paper/69d8843d7d0172140ec77c2a>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-08-01T05:00:05.813Z

Authors: Wenli Zhang, Xianglong Shi, Sirui Zhao, Xinqi Chen, Guo Cheng, Yifan Xu, Tong Xu, Yong Liao

Published: 2026-04-09T16:03:24.000Z

## 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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Talking-head Generation on CelebA-HQ paired with LibriSpeech (test)](<https://www.sota2.com/research/sota/talking-head-generation-on-celeba-hq-paired-with-librispeech-test.md>): Talking Head Generation; CelebA-HQ paired with LibriSpeech (test); V-PSNR: 22.76.

- [Talking Head Generation on HDTF (test)](<https://www.sota2.com/research/sota/talking-head-generation-on-hdtf-test.md>): Talking Head Generation; HDTF (test); FID: 3.69.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2604.08405>)

[PDF](<https://arxiv.org/pdf/2604.08405.pdf>)
