Reproduce “Instant Expressive Gaussian Head Avatars at Over 100 FPS” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kaiwen Jiang, Xueting Li, Seonwook Park, Ravi Ramamoorthi, Shalini De Mello, Koki Nagano
Instant4D animates a portrait from one image by encoding it as a reusable 3D Gaussian avatar and applying local, feature-space updates guided by a driving image. This avoids generating each frame with a diffusion model and supports viewpoint changes. On VOODOO-XP cross-reenactment, it reports 107.31 FPS and improves several expression and consistency metrics over X-NeMo, but ties InvertAvatar on MEt3R.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.