Handwriting Recognition System

This repository is the Tensorflow implementation of the Handwriting Recognition System described in Handwriting Recognition of Historical Documents with Few Labeled Data (please cite the paper if you use this code in your research paper). This code was also used for the baseline system in Fine-tuning Handwriting Recognition systems with Temporal Dropout.

This code is free for academic and research use. For commercial use of the code please contact Edgard Chammas.

To help run the system, sample images from ICDAR2017 Competition on Handwritten Text Recognition on the READ Dataset are added.

Configuration

General configuration can be found in config.py

CNN-specific architecture configuration can be found in cnn.py

Training

python train.py

This will generate a text log file and a Tensorflow summary.

Decoding

python test.py

This will generate, for each image, the line transcription. The output will be written to decoded.txt by default.

python compute_probs.py

This will generate, for each image, the posterior probabilities at each timestep. Files will be stored in Probs by default.

Dependencies

• Tensorflow
• OpenCV-Python

Citation

Please cite the following paper if you use this code in your research paper:

@inproceedings{chammas2018handwriting,
  title={Handwriting Recognition of Historical Documents with few labeled data},
  author={Chammas, Edgard and Mokbel, Chafic and Likforman-Sulem, Laurence},
  booktitle={2018 13th IAPR International Workshop on Document Analysis Systems (DAS)},
  pages={43--48},
  year={2018},
  organization={IEEE}
}

Acknowledgment

We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research.

Contributions

Feel free to send your pull request or open issues.