Reproduce “MobilePortrait: Real-Time One-Shot Neural Head Avatars on Mobile Devices” 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
Jianwen Jiang, Gaojie Lin, Zhengkun Rong, Chao Liang, Yongming Zhu, Jiaqi Yang, Tianyun Zhong
MobilePortrait animates a single portrait using motion from video or audio. It combines facial and neural keypoints to guide motion estimation, and supplies reusable appearance features to a lightweight synthesis network. In same-identity video reenactment, it uses 16 GFLOPs versus MCNet’s 200, with lower AKD but higher FID; this supports competitive video results, not universal quality or speed.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.