fendaq / textline-segmentation-using-fcn

Installations for running keras-theano on GPU

1. Upgrade pip and install opencv2

cd ~
pip install --upgrade pip
pip install opencv-python

2. Upgrade keras and set theano background

pip uninstall keras
pip install keras
vi ~/.keras/keras.json
  {
    "backend": "theano",
    "image_data_format": "channels_first",
    "floatx": "float32",
    "epsilon": 1e-07
  }

3. Upgrade theano

pip uninstall theano
pip install theano==1.0.1

4. Install pygpu which is necessary to run theano on GPU

git clone https://github.com/Theano/libgpuarray.git
cd libgpuarray
mkdir Build
cd Build
cmake .. -DCMAKE_BUILD_TYPE=Release
make
make install
cd ..
python3 setup.py build_ext -L /root/libgpuarray/lib -I /root/libgpuarray/include
python3 setup.py install
sudo ldconfig

5. Run python on a specific GPU

THEANO_FLAGS=device=cuda0 python3

How to run the project

1. Download challenging text line dataset

2. Preprocess

python3 preprocess.py

3. Train

THEANO_FLAGS=device=cuda0 python3

4. Predict

python3 SlidingPredictor.py

5. Postprocess

python3 directional_morphing.py