Launch Teddy icon

Scene Dreamer

No ratings
Generated 3D landscapes from 2D images.
Generated by ChatGPT

SceneDreamer is an unconditional generative model for unbounded 3D scenes that can synthesize large-scale 3D landscapes from random noises. The framework is learned from in-the-wild 2D image collections without any 3D annotations.

At the core of the tool is a principled learning paradigm comprising an efficient and expressive 3D scene representation, a generative scene parameterization, and an effective renderer that leverages the knowledge from 2D images.

SceneDreamer employs an efficient bird's-eye-view (BEV) representation generated from simplex noise, which consists of a height field and a semantic field.

The height field represents the surface elevation of 3D scenes, while the semantic field provides detailed scene semantics. The BEV scene representation enables the tool to represent a 3D scene with quadratic complexity, disentangle geometry and semantics, and perform efficient training.

The tool proposes a novel generative neural hash grid to parameterize the latent space given 3D positions and the scene semantics, which aims to encode generalizable features across scenes and align content.

Lastly, a neural volumetric renderer, learned from 2D image collections through adversarial training, is employed to produce photorealistic images. SceneDreamer is effective in generating vivid and diverse unbounded 3D worlds and is superior to state-of-the-art methods in this regard.

Scene Dreamer was manually vetted by our editorial team and was first featured on April 9th 2023.
Featured banner
Promote this AI Claim this AI

Would you recommend Scene Dreamer?

Help other people by letting them know if this AI was useful.


Pros and Cons


Generates 3D from 2D
Unbounded 3D scenes
Synthesizes from random noises
No need 3D annotations
Efficient 3D scene representation
Generative scene parameterization
Effective renderer
Bird's-eye-view representation
Utilizes simplex noise
Separate height and semantic field
3D scene with quadratic complexity
Disentangle geometry and semantics
Generative neural hash grid
Encodes generalizable features
Neural volumetric renderer
Photorealistic image production
Superior to state-of-the-art
Generates diverse 3D worlds
Free camera trajectory
Detailed scene semantics
Efficient training
Style modulated renderer
Can blend latent features
End-to-end training
Scene consistency and depth


Doesn't support 3D annotations
Only trained on 2D images
Limited surface elevation representation
Largely dependent on 2D semantics
Might not generalize well across different scene styles
No explicit depth handling
Fixed camera trajectory
Limited expressiveness in 3D representation
Dependent on adversarial training
Complexity due to quadratic representations


What is SceneDreamer?
How does SceneDreamer generate 3D scenes from 2D images?
What technologies underpin SceneDreamer's functionality?
What is the BEV scene representation in SceneDreamer?
What purpose does the height field serve in SceneDreamer?
What's the role of the semantic field in SceneDreamer?
How does SceneDreamer handle the complexity of 3D scenes?
Can you explain in simple terms how SceneDreamer uses a generative neural hash grid?
What is the neural volumetric renderer in SceneDreamer?
How does SceneDreamer use adversarial training?
What kind of images can SceneDreamer create?
How does SceneDreamer compare to other AI tools for scene generation?
Can SceneDreamer generate any type of landscape?
How diverse are the 3D worlds that SceneDreamer can create?
What does the term 'unbounded 3D scene' mean in the context of SceneDreamer?
Who created SceneDreamer?
Is the source code for SceneDreamer available?
Is there a video demonstration of how SceneDreamer works?
What's the significance of the style code in SceneDreamer?
What kind of 2D image collections can be used with SceneDreamer?


+ D bookmark this site for future reference
+ ↑/↓ go to top/bottom
+ ←/→ sort chronologically/alphabetically
↑↓←→ navigation
Enter open selected entry in new tab
⇧ + Enter open selected entry in new tab
⇧ + ↑/↓ expand/collapse list
/ focus search
Esc remove focus from search
A-Z go to letter (when A-Z sorting is enabled)
+ submit an entry
? toggle help menu
0 AIs selected
Clear selection