Dream-in-4D: A Unified Approach for Text- and Image-guided 4D Scene Generation

CVPR 2024

Paper Project Code

Additional Ablation Results

We compare with ablation baselines via a user preference study in the paper. Qualitative results are shown below.

Hexplane

w/o 2D diffusion prior

w/o deformation reg.

w/o multi-res. features

Ours

3D rendering of a fox playing videogame.

A baby panda eating ice cream.

A cat singing.

A dog riding a skateboard.

A goat drinking beer.

A monkey eating a candy bar.

An emoji of a baby panda reading a book.

A squirrel riding a motorcycle.

Clown fish swimming through the coral reef.

Superhero dog with red cape flying through the sky.