4DGen: Grounded 4D Content Generation with Spatial-temporal

gen4dRecent developments in 2D visual generation have been remarkably successful. However, 3D and 4D generation remain challenging in real-world applications due to the lack of large-scaleAided by text-to-image and text-to-video diffusion models, existing 4D content creation pipes utilize score distillation sampling to optimize the entire dynamic 3D scene.