EDTalk++ disentangles talking-head video into four orthogonal motion banks (mouth, pose, eyes, expression) and drives them from either video or audio inputs.
Difftalk: Crafting diffusion models for generalized audio-driven portraits anima- tion,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
EDTalk++: Full Disentanglement for Controllable Talking Head Synthesis
EDTalk++ disentangles talking-head video into four orthogonal motion banks (mouth, pose, eyes, expression) and drives them from either video or audio inputs.