GRMM combines a classic 3D face template with learned fine detail residuals to render controllable full-head avatars in real time.
Their combination enables robust learning of geometry, appearance, and expression disentanglement
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.GR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
GRMM: Real-Time High-Fidelity Gaussian Morphable Head Model with Learned Residuals
GRMM combines a classic 3D face template with learned fine detail residuals to render controllable full-head avatars in real time.