https://github.com/pvnieo/mangan

Automatic Manga colorization using Generative Adversarial Networks.
https://github.com/pvnieo/mangan

adversarial-networks automatic-colorization deep-learning generative-adversarial-network manga python3 pytorch

Last synced: about 1 year ago
JSON representation

Automatic Manga colorization using Generative Adversarial Networks.

• Host: GitHub
• URL: https://github.com/pvnieo/mangan
• Owner: pvnieo
• License: mit
• Created: 2018-12-31T18:07:36.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2019-04-10T19:01:28.000Z (about 7 years ago)
• Last Synced: 2025-03-26T21:47:38.824Z (over 1 year ago)
• Topics: adversarial-networks, automatic-colorization, deep-learning, generative-adversarial-network, manga, python3, pytorch
• Language: Python
• Size: 275 KB
• Stars: 19
• Watchers: 2
• Forks: 6
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# ManGAN

Automatic Manga colorization using Generative Adversarial Networks.

This project aims to color manga using GANs. For the moment, two types of GANs have been tested: [pix2pix](https://arxiv.org/pdf/1611.07004.pdf) and [cycle-gan](https://arxiv.org/pdf/1703.10593.pdf).

In this repo, we implemented cycle-gan in pytorch.

**Project Report:** [here](https://www.researchgate.net/publication/331175026_Manga_Colorization_using_Generative_Adversarial_Nets)

Some of the results obtained are displayed in the following image:

![Rusults](/assets/exp2.jpeg)

This project is **under development**, currently, there are diffuclties to color images containing text.

Installation instructions will be updated soon as well as our learned weights for the Death Note anime.