Reproduce “EmoTaG: Emotion-Aware Talking Head Synthesis on Gaussian Splatting with Few-Shot Personalization” 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

Haolan Xu, Keli Cheng, Lei Wang, Ning Bi, Xiaoming Liu

EmoTaG animates personalized 3D talking-head avatars from brief video, using audio plus pose and upper-face expression cues. It predicts expression and jaw motion in FLAME, then transfers motion to face-attached Gaussians. A gated network adds emotional movement to shared speech articulation. In five-second emotional self-reconstruction, it achieves PSNR 29.95 versus 27.82 for InsTaG, but is not best on Sync-C.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.