Skip to content

LarsSluijter/SingleTerrainGAN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

SingleTerrainGAN

SingleTerrainGAN is a GAN that can generate terrains following two datasets fetched from studies by Beckham and Pal (2017) and Panagiotou and Charou (2020). SingleTerrainGAN is a GAN trained for the research "Improving procedural terrain generation using a single deep learning model".

What's in this repository?

This repository includes the following files/directories:

  • Notebooks/Gan-heightmaps-recreation.ipynb - Python notebook that contains a recreation of the GAN developed by Beckham and Pal in TensorFlow.
  • Thesis.pdf - Full thesis of the research.
  • Survey data.csv - Data obtained from a survey carried out for the research.
  • Survey models - 3D models used for the survey

How to generate your own terrains

To generate your own generates, you will first have to downloaded the trained model in this Google Drive folder. This folder includes two different models for SingleterrainGAN: "SingleTerrainGAN_desert.pkl" and "SingleTerrainGAN_Greece.pkl". You can download a model according to the type of terrain (desert/Greece) you would like to generate. You can then use this .pkl file to generate terrains using the code in the stylegan2-fun repository by Diego Porres.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published