# OCR-tf2

**Repository Path**: darkyu/OCR-tf2

## Basic Information

- **Project Name**: OCR-tf2
- **Description**: this project implements text area detection and OCR
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-05-28
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# OCR-tf2
this project implements text area detection and OCR

## download the dataset for text area detection

download the dataset prepared by the author of the paper "Detecting Text in Natural Image with Connectionist Text Proposal Network" [here](https://pan.baidu.com/s/1nbbCZwlHdgAI20_P9uw9LQ)

## create dataset for text area detection

create with the following command

```bash
python3 create_dataset.py <path/to/mlt directory>
```

## train the text area detector

train with the following command

```bash
python3 train.py ctpn
```

when the training process finishes, it will save the parameters in hdf5 format automatically. if you interrupt the training process, you can save the model with the following command

```bash
python3 save_model.py ctpn
```

## test the text area detector

test the detector with

```bash
python3 TextDetector.py <path/to/the/picture>
```

## text area detection results

here are some results of my model which is enclosed at model/ctpn.h5 .

<p align="center">
  <table>
    <caption>Loss</caption>
    <tr><td><img src="pics/ctpn/loss.png" alt="train loss" width="800" /></td></tr>
  </table>
</p>
<p align="center">
  <table>
    <caption>Detection results</caption>
    <tr>
      <td><img src="pics/ctpn/result1.png" width="400" /></td>
      <td><img src="pics/ctpn/result2.png" width="400" /></td>
    </tr>
    <tr>
      <td><img src="pics/ctpn/result3.png" width="400" /></td>
      <td><img src="pics/ctpn/result4.png" width="400" /></td>
    </tr>
  </table>
</p>

## train the ocr

the OCR part of the project is an implement of CRNN introduced in paper "An End-to-End Trainable Neural Network for Image-based Sequence Recognition and Its Application to Scene Text Recognition" . train the model with the following command

```bash
python3 train.py ocr
```

save the trained model with command

```bash
python3 save_model.py ocr
```