https://github.com/kvfrans/deepcolor

Automatic coloring and shading of manga-style lineart, using Tensorflow + cGANs
https://github.com/kvfrans/deepcolor

Last synced: 3 months ago
JSON representation

Automatic coloring and shading of manga-style lineart, using Tensorflow + cGANs

• Host: GitHub
• URL: https://github.com/kvfrans/deepcolor
• Owner: kvfrans
• Created: 2017-02-03T05:03:42.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2018-05-25T03:52:14.000Z (over 6 years ago)
• Last Synced: 2024-06-16T00:37:28.913Z (5 months ago)
• Language: Python
• Homepage: http://color.kvfrans.com
• Size: 65.5 MB
• Stars: 553
• Watchers: 21
• Forks: 86
• Open Issues: 11
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-list - deepcolor - Automatic coloring and shading of manga-style lineart, using Tensorflow + cGANs. (Computer Vision / Image / Video Generation)

README

        
# deepcolor

Automatic coloring and shading of manga-style lineart, using Tensorflow + cGANs

![](http://kvfrans.com/content/images/2017/03/Screen-Shot-2017-03-01-at-11-09-09-PM-1.png)

![](http://kvfrans.com/content/images/2017/03/Screen-Shot-2017-03-01-at-11-09-13-PM.png)

Read the writeup:

http://kvfrans.com/coloring-and-shading-line-art-automatically-through-conditional-gans/

Try the demo:

http://color.kvfrans.com

## Setup

### Prereqs

    - Python 2.7, numpy

    - Tensorflow 0.12

    - OpenCV

### Running it

1. make a folder called "results"

2. make a folder called "imgs"

3. Fill the "imgs" folder with your own .jpg images, or run "download_images.py" to download from Safebooru.

4. Run "python main.py train". I trained for ~20 epochs, taking about 16 hours on one GPU.

5. To sample, run "python main.py sample"

6. To start the server, run "python server.py". It will host on port 8000.

### Pre-trained

Get the pretrained model:

https://drive.google.com/file/d/0BydPPLNieijIdDlUYWxhelEwRnM/view?usp=sharing

Folder structure should go:

```

main.py

server.py

checkpoint/

    tr/

        checkpoint

        model-1101500.index

        model-1101500.data-00000-of-00001

        model-1101500.meta

```

Code based off [this pix2pix implementation](https://github.com/yenchenlin/pix2pix-tensorflow).