- How to identify fonts in pictures using deep learning? we can do visual font recognition (VFR) using DeepFont algorithm (from Adobe). The paper is here
![res](https://user-images.githubusercontent.com/8428372/205088069-94c51227-4870-4d87-bd0f-52ab16950715.png)
- This implementation is from Font_Recognition-DeepFont, in this repo
requirements.txt
and Dockerfile
are added so we can set it up easily
- Github action is added so we can make sure Dockerfile do work
- Data generation script
create_data.py
is provided
train.py
and eval.py
are provided so we can run training and evaluation without jupyter notebook
python3 -m "virtualenv" venv
source venv/bin/activate
pip3 install -r requirements.txt
- create dataset
-c
: the number of generated images for each font in fonts
folder
-f
: the target folder
python3 create_data.py -c 1 -f data
- evaluate
-d
: the path of testing file
-m
: model filename
python3 eval.py -d test_data/Times.jpg -m top_model.h5
docker build -t mydeepfont .
- Run dockerfile
- access
http://127.0.0.1:8888
by browser
docker run -it --rm -p8888:8888 mydeepfont