Tuning-free approaches adapting large-scale pre-trained video diffusion models for identity-preserving text-to-video generation (IPT2V) have gained popularity recently due to their efficacy and scalability. However, significant challenges remain to achieve satisfied facial dynamics while keeping the identity unchanged. In this work, we present a novel tuning-free IPT2V framework by enhancing face knowledge of the pre-trained video model built on diffusion transformers (DiT), dubbed FantasyID. Essentially, 3D facial geometry prior is incorporated to ensure plausible facial structures during video synthesis. To prevent the model from learning ``copy-paste'' shortcuts that simply replicate reference face across frames, a multi-view face augmentation strategy is devised to capture diverse 2D facial appearance features, hence increasing the dynamics over the facial expressions and head poses. Additionally, after blending the 2D and 3D features as guidance, instead of naively employing adapter to inject guidance cues into DiT layers, a learnable layer-aware adaptive mechanism is employed to selectively inject the fused features into each individual DiT layers, facilitating balanced modeling of identity preservation and motion dynamics. Experimental results validate our model’s superiority over the current tuning-free IPT2V methods.
Ours model and code will be open sourced before May.
@misc{zhang2025fantasyidfaceknowledgeenhanced,
title={FantasyID: Face Knowledge Enhanced ID-Preserving Video Generation},
author={Yunpeng Zhang and Qiang Wang and Fan Jiang and Yaqi Fan and Mu Xu and Yonggang Qi},
year={2025},
eprint={2502.13995},
archivePrefix={arXiv},
primaryClass={cs.GR},
url={https://arxiv.org/abs/2502.13995},
}