An open API service indexing awesome lists of open source software.

https://github.com/jiangnanboy/table_structure_recognition

利用Swin-Unet(Swin Transformer Unet)实现对文档图片里表格结构的识别,Swin-unet (Swin Transformer Unet) is used to identify the document table structure
https://github.com/jiangnanboy/table_structure_recognition

swin-transformer-unet swin-unet table-recognition table-structure-recognition

Last synced: 5 months ago
JSON representation

利用Swin-Unet(Swin Transformer Unet)实现对文档图片里表格结构的识别,Swin-unet (Swin Transformer Unet) is used to identify the document table structure

Awesome Lists containing this project

README

          

[简体中文](README.md) | English
# Swin-unet (Swin Transformer Unet) is used to identify the document table structure
![](img/architecture.png)
## function
- [x] Identify the lines in the doc table
- [ ] Convert results to excel

## download the weights model file
Model file address:https://pan.baidu.com/s/1K2bxqqmZaOtPI1jeWN3Ygg

password:w0yj

Copy the model.h5 to the model directory

## training(tensorflow2.5)
See train.py in this project

## prediction
See table_line_infer.py in this project

## identification result presentation
![](img/2.jpg)
### table line detection
![](img/2_line.png)

![](img/3.jpg)
### table line detection
![](img/3_line.png)

![](img/8.jpg)
### table line detection
![](img/8_line.png)

![](img/dd.jpg)
### table line detection
![](img/dd_line.png)

![](img/plane.png)
### table line detection
![](img/plane_line.png)

![](img/table.png)
### table line detection
![](img/table_line.png)

## contact

1、github:https://github.com/jiangnanboy

2、blog:https://www.cnblogs.com/little-horse/

3、QQ:2229029156