1 paper
Lukas Höllein, Matthias Nießner
Video diffusion models generate high-quality and diverse worlds; however, individual frames often lack 3D consistency across the output sequence, which makes the reconstruction of…