OCRs text. Intended for use with manga pages, but could probably be extended to other uses.
Note this is not complete! As such, it won't work really well.
Note: This was initially made for https://guya.moe, and due to this tool being a bit of a pain to use, I settled on working with Azure via KaguyaOCR instead. See that instead though that's also going to get deprecated for another tool I'm working on.
- Create a venv
- Make, which will download the pip reqs.
- Run using
python3 image_localization.py
or just./image_localization.py
. Look at the command line options for what you can pass in.