TDN

[Terrain Diffusion Network: Climatic-Aware Terrain Generation with Geological Sketch Guidance] GitHub

[Terrain Diffusion Network] is a sketch to terrain multi-denoiser diffusion model, which actively incorporates user guidance for enhanced controllability, taking into account terrain features like rivers, ridges, basins, and peaks. Instead of adhering to a conventional monolithic denoising process, which often compromises the fidelity of terrain details or the alignment with user control, a multi-level denoising scheme is proposed to generate more realistic terrains by taking into account finegrained details, particularly those related to climatic patterns influenced by erosion and tectonic activities.

Set Up

conda create -n tdn python=3.9
conda activate tdn
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
pip install -r requirements.txt

Terrain Diffusion Network

Current version of Terrain Diffusion Network consists of three level denoiser with a sketch encoder and a terrain encoder, showed as below

Text-to-Image with Stable Diffusion

Inference Script

We provide a sample inference sampling script:

python test.py
 

Todo

• Upload training dataset to google drive
• Upload model

Dataset

https://drive.google.com/file/d/1P4OZXBIIRI_GRvyKNebV3bk42ciCAXqN/view?usp=sharing

BibTeX