Monocular and Generalizable Gaussian Talking Head Animation

Shengjie Gong Haojie Li Jiapeng Tang Dongming Hu Shuangping Huang, Hao Chen Tianshui Chen Zhuoman Liu · 2025

Read Monocular and Generalizable Gaussian Talking Head Animation by Shengjie Gong Haojie Li Jiapeng Tang Dongming Hu Shuangping Huang, Hao Chen Tianshui Chen Zhuoman Liu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-00665.txt.

Description

Gong leads a team presenting MGGTalk, a system that animates talking heads from a single video in a 2025 paper. Earlier Gaussian-based methods needed many camera views or per-person retraining. One ordinary clip is far easier to get.

Depth estimates supply geometry for the unseen parts of the face. Symmetry operations and point cloud filtering complete the head's position layout. A two-stage scheme then predicts the remaining Gaussian parameters, first for visible regions, then for hidden ones guided by the visible side. Tests beat prior state-of-the-art systems across several metrics. No personalized retraining is needed for new identities. The authors target practical uses like virtual presenters. Single-camera input is the selling point.

Catalog record

CREATORShengjie Gong Haojie Li Jiapeng Tang Dongming Hu Shuangping Huang, Hao Chen Tianshui Chen Zhuoman Liu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-00665

Browse the full library · About the Tower of Babel