Style Classification into Machine Printed and Handwritten Text

Keras implementation of Faster R-CNN to classify text into Machine Printed and Handwritten Text

Getting Started

Prerequisites

• numpy

$ pip install numpy

• h5py

$ pip install h5py

• opencv-python

$ pip install opencv-python

• sklearn

$ pip install scikit-learn

• Keras==2.0.3 (Both theano and tensorflow backends are supported. However tensorflow is recommended)

$ pip install Keras==2.0.3

• Tensorflow

$ pip install Tensorflow

Trained Model

Trained model can be downloaded from: https://drive.google.com/drive/folders/1eymjiH7_oWJbSI4LOAVhMzzNFtkbYcPz?usp=sharing

How to run Training:

• Copy pretrained weights for resnet50 (resnet50_weights_tf_dim_ordering_tf_kernels.h5) in Style-Classification directory.

• train_frcnn.py can be used to train a model. To train the data, it must be in PASCAL VOC format. To train simply do:

$ python train_frcnn.py -p /path/to/train_data/

• Running train_frcnn.py will write weights to disk to an hdf5 file, as well as all the setting of the training run to a pickle file. These settings can then be loaded by test_frcnn.py for any testing.

How to run Testing:

• Copy trained model(model_frcnn.hd5) and config.pickle file in Style-Classification diectory.

• test_frcnn.py can be used to perform inference, given pretrained weights and a config file. Specify a path to the folder containing images:

$ python test_frcnn.py -p /path/to/test_data/

CREDITS

This code is inspired from https://github.com/yhenon/keras-frcnn