https://github.com/tuandoan998/handwritten-text-recognition

IAM dataset
https://github.com/tuandoan998/handwritten-text-recognition

cnn iam keras ocr python3 rnn

Last synced: 3 months ago
JSON representation

IAM dataset

• Host: GitHub
• URL: https://github.com/tuandoan998/handwritten-text-recognition
• Owner: tuandoan998
• Created: 2019-08-25T14:39:52.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2023-03-22T08:11:19.000Z (about 2 years ago)
• Last Synced: 2025-03-18T17:51:30.542Z (3 months ago)
• Topics: cnn, iam, keras, ocr, python3, rnn
• Language: Python
• Homepage:
• Size: 76.9 MB
• Stars: 56
• Watchers: 2
• Forks: 16
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# OCR

This project refers to the handwritten recognition with CNN and RNN, decode with CTC.

![demo](https://github.com/tuandoan998/OCR_IAM-dataset/blob/master/Resource/demo.png)

[Demo on youtube](https://youtu.be/kILhJXcR7To)

# Update

This [repo](https://github.com/tuandoan998/vietnamese-htr) has better model and higher accuracy 

## Dataset

[IAM dataset](http://www.fki.inf.unibe.ch/databases/iam-handwriting-database/download-the-iam-handwriting-database)  

* [Model1 - word_model.png] Train on WORD unit of dataset.

* [Model2 - line_model.png] Train on LINE unit of dataset.

* Split data:

https://www.openslr.org/56/

## Result

Test on IAM dataset:

|  Model  | Test Unit | CER(%) | WER(%) | 

| :-      | :-        |  ---:  |  ---:  |

|  WORD   | WORD      | 17.20  | 37.78  | 

## Train

[Google colab]

## Usage

### Training

```

$ python3 Train.py

```

### Predict

```

$ python3 Prediction.py

```

![predict](https://github.com/tuandoan998/OCR_IAM-dataset/blob/master/Resource/predict.png)

### Evaluation

```

$ python3 EvaluationWord.py

$ python3 EvaluationLine.py

```

## References

https://towardsdatascience.com/build-a-handwritten-text-recognition-system-using-tensorflow-2326a3487cd5

http://norvig.com/spell-correct.html