A simple yet effective approach to rectify distorted document image by estimating control points and reference points.
The control points and reference points are composed of the same number of vertices and describe the shape of the document in the image before and after rectifying, respectively. The control points are controllable to facilitate interaction or subsequent adjustment. You can flexibly select post-processing methods and the number of vertices according to different application scenarios.
See “Document Dewarping with Control Points” for more information.
-
Test
python test.py --data_path_test=./your/test/data/path/ -
Train
python train.py --data_path_train=./your/train/data/path/ --data_path_validate=./your/validate/data/path/ --data_path_test=./your/test/data/path/ --batch_size 32 --schema train --parallel 01
python >=3.7
pytorch
opencv-python
scipy
The training dataset can be synthesised using the scripts.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent