Use of GAN approach in order to generate monte carlo noise applied on synthesis images

Jérôme BUISINE 714bc0055b Merge branch 'release/v0.0.2' 5 lat temu
synthesis_images 6901223dae Update of data; Save and load option added 5 lat temu
.gitignore a003b62672 Update of documentation 5 lat temu
Dockerfile b89725ffd3 Add of gan for image synthesis 5 lat temu
LICENSE a003b62672 Update of documentation 5 lat temu
Makefile b89725ffd3 Add of gan for image synthesis 5 lat temu
README.md a003b62672 Update of documentation 5 lat temu
ganAtariImage.py b89725ffd3 Add of gan for image synthesis 5 lat temu
ganSynthesisImage.py 70d4356303 try using 200 pixels images 5 lat temu
ganSynthesisImage_100.py 51f1999d4f Add of gan for 100x100 images 5 lat temu
ganSynthesisImage_200.py 6901223dae Update of data; Save and load option added 5 lat temu
noise_gan.ipynb 51f1999d4f Add of gan for 100x100 images 5 lat temu
prepare_data.py 6901223dae Update of data; Save and load option added 5 lat temu
requirements.txt b89725ffd3 Add of gan for image synthesis 5 lat temu
tensorboard.ipynb 51f1999d4f Add of gan for 100x100 images 5 lat temu

README.md

Noise Generation

Description

Study of how to generate noise filter from monte carlo rendering in synthesis images using Generative Adversarial Network approach. The aim of this project is to reproduce monte carlo noise obtained during rendering.

How to use ?

pip install -r requirements.txt
python prepare_data.py
python ganSynthesisImage_200.py

Contributors

License

MIT