Street View
Videos of novel scenes created by SRT on the street view dataset. For each scene, 5 input views are provided to the model. The scene representation is computed in a forward pass in ~10ms. 96 frames are rendered with ~120 fps.
Videos of novel scenes created by SRT on the street view dataset. For each scene, 5 input views are provided to the model. The scene representation is computed in a forward pass in ~10ms. 96 frames are rendered with ~120 fps.