# tf_doc_localisation

**Repository Path**: tunmx/tf_doc_localisation

## Basic Information

- **Project Name**: tf_doc_localisation
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

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

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# tf_doc_localisation

### Requirements

1. tensorflow version==1.12.0
2. PIL (pip install PIL)
3. download [part of the dataset](https://drive.google.com/drive/u/0/folders/1D7tv5NkFlnVWZQ3ViIO2EHpYK0nlGRfV)，extract it to $MIDV_DATASET
4. Edit tf_doc_localisation/receipt_dataset.py line 18 

``` python
midv_dir = MIDV_DATASET ## for example: /data/your_download_directory/midv_500
```

### how to train the network

1. cd tf_doc_localisation
2. mkdir data
3. python synthesis_data.py
4. python receipt_dataset.py
5. python train.py

### how to test

1. python test.py


some test samples.

#### case 1

![case1](https://raw.githubusercontent.com/RRanddom/tf_doc_localisation/master/raw_data/demo_images/case1.png)

#### case 2

![case2](https://raw.githubusercontent.com/RRanddom/tf_doc_localisation/master/raw_data/demo_images/case2.png)

#### failure case

![failure case](https://raw.githubusercontent.com/RRanddom/tf_doc_localisation/master/raw_data/demo_images/case3.png)

### todo

1. add evaluation code.
2. 🚧 🚧 🚧 🚧 🚧