https://github.com/0x454447415244/HandwritingRecognitionSystem

Handwriting Recognition System based on a deep Convolutional Recurrent Neural Network architecture
https://github.com/0x454447415244/HandwritingRecognitionSystem

cnn deep-learning handwriting-recognition machine-learning rnn tensorflow

Last synced: about 2 months ago
JSON representation

Handwriting Recognition System based on a deep Convolutional Recurrent Neural Network architecture

• Host: GitHub
• URL: https://github.com/0x454447415244/HandwritingRecognitionSystem
• Owner: 0x454447415244
• License: other
• Created: 2018-10-27T02:19:21.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2023-08-22T01:05:09.000Z (about 1 year ago)
• Last Synced: 2024-06-10T11:47:54.290Z (3 months ago)
• Topics: cnn, deep-learning, handwriting-recognition, machine-learning, rnn, tensorflow
• Language: Python
• Size: 4.9 MB
• Stars: 415
• Watchers: 22
• Forks: 136
• Open Issues: 5
• Metadata Files:
  • Readme: README.md
  • License: License.md

Awesome Lists containing this project

README

        
# 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is the Tensorflow implementation of the Handwriting Recognition System described in [Handwriting Recognition of Historical Documents with Few Labeled Data](https://www.researchgate.net/publication/325993975_Handwriting_Recognition_of_Historical_Documents_with_Few_Labeled_Data) (please cite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](https://www.researchgate.net/publication/348958179_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](mailto:[email protected]).

To help run the system, sample images from [ICDAR2017 Competition on Handwritten Text Recognition on the READ Dataset](https://scriptnet.iit.demokritos.gr/competitions/8/) 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.