FactorPortrait: Controllable Portrait Animation via
Disentangled Expression, Pose, and Viewpoint

1Meta Reality Labs 2Technical University of Munich

Phone Dataset

It is a monocular iPhone video dataset from a fixed frontal view, including a variety of actions such as head rotation, brief expressions, and speech. In this dataset, we want to generate static frontal view videos with changes in pose and expression.


Self Reenactment

Input GAGA CAP4D HunyuanPortrait Ours GT
Input GAGA CAP4D HunyuanPortrait Ours GT

Cross Reenactment

Input GAGA CAP4D HunyuanPortrait Ours Reference/Driving
Input GAGA CAP4D HunyuanPortrait Ours Reference/Driving

Page Title