Customizing Video Portraits via Identity-ActionDecouplingCustomizing Video Portraits via Identity-ActionDecoupling

arXiv2026

Junxiong Lin, Haoran Wang, Xinji Mai, Zeng Tao, Xuan Tong, Ivy Pan, Wenqiang Zhang

Figure from the 图片来源：original paper原论文 ↗

IaD separates identity and facial action features to generate identity-consistent videos with text-controlled expressions, without subject-specific fine-tuning.

IaD 分离身份与面部动作特征，在无需针对特定人物微调的情况下，生成身份一致且表情受文本控制的人像视频。