anopara/texture-synthesis-nonparametric-sampling
Implementation of “Texture Synthesis with Non-Parametric Sampling” paper by Alexei A. Efros and Thomas K. Leung
| repo name | anopara/texture-synthesis-nonparametric-sampling |
| repo link | https://github.com/anopara/texture-synthesis-nonparametric-sampling |
| homepage | |
| language | Jupyter Notebook |
| size (curr.) | 5530 kB |
| stars (curr.) | 108 |
| created | 2018-12-05 |
| license | MIT License |
Texture Synthesis
Based on “Texture Synthesis with Non-Parametric Sampling” paper by Alexei A. Efros and Thomas K. Leung
There are 2 versions: Python and Houdini
Python
For Python version you would need:
- Python 3.7
- Jupyter Notebook (my version is 5.6.0)
- Numpy (my version is 1.15.1)
- Matplotlib (2.2.3)
- Scipy (1.1.0)
- Skimage (0.14.0)
- imageio (2.4.1) (if you want to make a GIF :))
- PIL (5.2.0)
To start, open the Jupyter Notebook file “Texture Synthesis”, and follow the instructions :)
Houdini
For Houdini:
- Houdini 17 (might work with earlier versions, haven’t tried)
To start, open TextureSynthesis<…>.hipnc file and follow the instructions inside (they are scarce, but present) Use this file at your own risk :D WARNING: it’s very very slow! (you might wanna use it on smaller images…) and the results are less nice that the Python version
