Skip to content

Genues/AngelinaReader

 
 

Repository files navigation

Angelina braille reader

Draft version. Production version will be at https://github.com/IlyaOvodov/AngelinaReader (under construction)

requirements

ubuntu, windows with GPU

CUDA
Python 3.6
PyTorch 1.4
torchvision
ignite
numpy
PIL
albumentations
cv2
https://github.com/IlyaOvodov/pytorch-retinanet
https://github.com/IlyaOvodov/OvoTools
https://github.com/IlyaOvodov/labelme (for annotation)

installation

git clone https://github.com/IlyaOvodov/BrailleV0.git
edit BrailleV0/local_config.py to set data_path and global_3rd_party pointing to current dir
git clone https://github.com/IlyaOvodov/OvoTools.git
cd OvoTools
python setup.py develop
cd ..
git clone https://github.com/IlyaOvodov/pytorch-retinanet.git pytorch_retinanet
download model https://yadi.sk/d/GW0qEmA5rL0m0A into ./NN_saved/retina_chars_eced60/models/clr.008

usage

cd BrailleV0/web-app
python angelina_reader_app.py
access it by 127.0.0.1:5000

or edit if __name__=="__main__" section BrailleV0/NN/RetinaNet/infer_retinanet.py. Set img_filename_mask and results_dir to proper values and run it.

About

Angelina Braille Reader. Braille photo -> text.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 86.6%
  • HTML 13.3%
  • Batchfile 0.1%