Models from paper M. Kišš, M. Hradiš, and O. Kodym, “Brno Mobile OCR Dataset” in 2019 15th IAPR International Conference on Document Analysis and Recognition (ICDAR), IEEE, 2019. (arxiv)
We also provide a python script for transcribing text lines with our proposed networks. The script runs with parameters specifying directory with image to transcribe and also used network. For example
python3 transcribe.py --input-dir=/tmp/my_dir/ --net=LSTM
transcribes all .jpg
and .png
files in /tmp/my_dir/
using neural network with recurrent layers (network stored in lstm_net/
).
If you want to transcribe lines using netowrk without recurrent layers, use --net=CONV
parameter instead.