1 paper
Xinan He, Kaiqing Lin, Yue Zhou +8
With the rapid advancement of video generation models such as Veo and Wan, the visual quality of synthetic content has reached a level where macro-level semantic errors and tempora…