https://github.com/rapidai/rapidundistort
修正文档扭曲/模糊/阴影等情况，使用onnx模型简单轻量部署，未来持续跟进最新最好的文档矫正方案和模型,Correct document distortion using a lightweight ONNX model for easy deployment. We will continue to follow and integrate the latest and best document correction solutions and models in the future.
https://github.com/rapidai/rapidundistort
Last synced: 10 months ago
JSON representation
修正文档扭曲/模糊/阴影等情况，使用onnx模型简单轻量部署，未来持续跟进最新最好的文档矫正方案和模型,Correct document distortion using a lightweight ONNX model for easy deployment. We will continue to follow and integrate the latest and best document correction solutions and models in the future.
- Host: GitHub
- URL: https://github.com/rapidai/rapidundistort
- Owner: RapidAI
- Created: 2024-11-05T05:32:22.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-15T09:16:30.000Z (over 1 year ago)
- Last Synced: 2025-06-07T22:06:19.341Z (about 1 year ago)
- Language: Python
- Homepage:
- Size: 34.1 MB
- Stars: 56
- Watchers: 3
- Forks: 9
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
📊RapidUndistort
[English](README_en.md) | 简体中文
### 最近更新
- **2024.11.15**
- 完成初版代码，转换 [UVDoc](https://github.com/tanguymagne/UVDoc) 模型为onnx,完善前后处理
- **2024.12.15**
- 补充去模糊/去阴影/二值化的功能和模型，重新升级为 RapidUndistort
### 简介
本仓库用于进行文档扭曲的修正/文档去模糊/文档去阴影/文档二值化等问题 \
提供多种模型和自由组合选择task，支持模型自动下载 \
原始pytorch模型来源参考 [致谢](#致谢) \
[快速使用](#快速使用) [使用建议](#使用建议) [参数说明](#参数说明) [模型地址](https://www.modelscope.cn/studios/jockerK/DocUnwrap/files)
### 在线体验
[modelscope](https://www.modelscope.cn/studios/jockerK/DocUnwrap) [huggingface](https://huggingface.co/spaces/Joker1212/RapidUnwrap)
### 效果展示
### 安装
``` python {linenos=table}
# 建议使用清华源安装 https://pypi.tuna.tsinghua.edu.cn/simple
pip install rapid-undistorted
```
### 快速使用
``` python {linenos=table}
import cv2
from rapid_undistorted.inference import InferenceEngine
img_path = "img/demo.jpg"
engine = InferenceEngine()
# 扭曲修正->去阴影->去模糊 (指定去模糊模型)
output_img, elapse = engine(img_path, ["unwrap", "unshadow", ("unblur", "OpenCvBilateral")])
# 去阴影->去模糊 (指定去模糊模型)
#output_img, elapse = engine(img_path, ["unshadow", ("unblur", "OpenCvBilateral")])
# 默认选择yaml配置文件中第一个unblur模型
#output_img, elapse = engine(img_path, ["unshadow", "unblur"])
# 二值化替代去阴影方法
#output_img, elapse = engine(img_path, ["unwrap", "binarize", "unblur"])
print(f"doc unwrap elapse:{elapse}")
cv2.imwrite("result.png", output_img)
```
### 使用建议
- 去模糊 NAFDPM 模型和直接使用opencv方法都有适用的情况,最好还是在线实际测试为准
- 去阴影模型相对于二值化功能更丰富，效果更好，不建议直接使用二值化方法
### 参数说明
#### 初始化参数
支持传入一个config配置文件，声明需要的task类型和对应的模型，以及model path
```python
config_path = "configs/config.yaml"
engine = InferenceEngine(config_path)
```
```yaml
tasks:
unwrap:
models:
- type: "UVDoc"
path:
use_cuda: false
unshadow:
models:
- type: "GCDnet"
sub_models:
- type: "GCDnet"
path:
use_cuda: false
use_dml: false
- type: "DRnet"
path:
use_cuda: false
binarize:
models:
- type: "UnetCnn"
path:
use_cuda: false
unblur:
models:
- type: "OpenCvBilateral"
path:
- type: "NAFDPM"
path:
use_cuda: false
```
#### 执行参数
```python
engine(img_path, task_list)
engine(img_path, ["unwrap", "unshadow", ("unblur", "OpenCvBilateral")])
```
### 致谢
unwrap: [UVDoc](https://github.com/tanguymagne/UVDoc)
unshadow: [GCDnet](https://github.com/ZZZHANG-jx/GCDRNet)
unblur: [NAFDPM](https://github.com/ispamm/NAF-DPM)
binarize: [UnetCnn](https://github.com/sajjanvsl/U-Net-CNN-for-binarization-of-Historical-Kannada-Handwritten-Palm-Leaf-Manuscripts)
### 贡献指南
欢迎提交请求。对于重大更改，请先打开issue讨论您想要改变的内容。
有其他的好建议和集成场景，作者也会积极响应支持
### 开源许可证
该项目采用[Apache 2.0](https://github.com/RapidAI/TableStructureRec/blob/c41bbd23898cb27a957ed962b0ffee3c74dfeff1/LICENSE)
开源许可证。