EvalTalker: Learning to Evaluate Real-Portrait-Driven Multi-Subject Talking Humans

Yingjie Zhou Xilei Zhu Siyu Ren Ziyi Zhao Ziwen Wang, Farong Wen Yu Zhou Jiezhang Cao Xiongkuo Min Fengjiao Chen, Xiaoyu Li Xuezhi Cao Guangtao Zhai Xiaohong Liu · 2025

Read EvalTalker: Learning to Evaluate Real-Portrait-Driven Multi-Subject Talking Humans by Yingjie Zhou Xilei Zhu Siyu Ren Ziyi Zhao Ziwen Wang, Farong Wen Yu Zhou Jiezhang Cao Xiongkuo Min Fengjiao Chen, Xiaoyu Li Xuezhi Cao Guangtao Zhai Xiaohong 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-2512-01340.txt.

Description

Yingjie Zhou and a large team work on judging fake talking faces. Speech-driven tools can make a still face seem to talk. New tools can move two or more faces at once. That adds life to calls and shows, but quality drops as faces are added.

Judging that quality by eye is slow and uneven. The team built EvalTalker to learn the judging task itself. It scores clips made from real face photos with more than one subject. Fair scores help builders spot flaws fast. They also give users a clear sign of which tools to trust.

Catalog record

CREATORYingjie Zhou Xilei Zhu Siyu Ren Ziyi Zhao Ziwen Wang, Farong Wen Yu Zhou Jiezhang Cao Xiongkuo Min Fengjiao Chen, Xiaoyu Li Xuezhi Cao Guangtao Zhai Xiaohong Liu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-01340

Browse the full library · About the Tower of Babel