1 paper
Zhexin Lou, Guancheng Lu, Zeyu Zhang +3
Pretrained world models can generate diverse environments, yet users often want to explore a particular scene specified by their own video. This requires learning the scene's visua…