EagleJumpSystem is a yonkoma manga management system.

Search koma(panel) with script and characters.

Automatic text detection by Google Cloud Vision API. Also, you could fix it by hands.

Half-automatic character detection by machine-learning. Once you annotate some pages(e.g. 100 for each characters), your machine annotate the rest pages.

Install Python3 and OpenCV. I recomend use pyenv and anaconda like following:

pyenv install anaconda3-2.4.0
pyenv virtualenv anaconda3-2.4.0 gochiusa
pyenv local anaconda3-2.4.0/envs/gochiusa
conda install -c https://conda.binstar.org/menpo opencv3

Obtain API key(server key) for Google Vision API, and save as /path/to/auth.json.

export GOOGLE_APPLICATION_CREDENTIALS="/path/to/auth.json"
pip install -r requirements.txt
npm install
npm run gulp build
python app/main.py

and open http://localhost:5000

All data is stored at ./data. So, use Dropbox.

mv ./data ~/Dropbox/
ln -s ~/Dropbox/ data