The code and the DIW dataset for "Learning From Documents in the Wild to Improve Document Unwarping" (SIGGRAPH 2022)

Go to file

SWHL 3647958e9c Update README (#12 )		2022-12-20 00:04:24 -05:00
configs	initial commit	2022-04-17 21:50:36 -04:00
data	add diw file list	2022-04-17 22:44:54 -04:00
eval	initial commit	2022-04-17 21:50:36 -04:00
images	Add the demo of inferring one image. (#7 )	2022-10-23 01:56:27 -04:00
networks	initial commit	2022-04-17 21:50:36 -04:00
utils	initial commit	2022-04-17 21:50:36 -04:00
.gitignore	Add requirements.txt, remove .vscode folder and add hugging face demo. (#8 )	2022-10-26 00:18:36 -04:00
LICENSE	Initial commit	2022-04-17 21:44:16 -04:00
README.md	Update README (#12 )	2022-12-20 00:04:24 -05:00
demo.py	Add the demo of inferring one image. (#7 )	2022-10-23 01:56:27 -04:00
eval.py	initial commit	2022-04-17 21:50:36 -04:00
eval.sh	initial commit	2022-04-17 21:50:36 -04:00
requirements.txt	Add requirements.txt, remove .vscode folder and add hugging face demo. (#8 )	2022-10-26 00:18:36 -04:00
train.py	initial commit	2022-04-17 21:50:36 -04:00
train.sh	initial commit	2022-04-17 21:50:36 -04:00

README.md

PaperEdge

The code and the DIW dataset for "Learning From Documents in the Wild to Improve Document Unwarping" (SIGGRAPH 2022)

[paper] [supplementary material]

Documents In the Wild (DIW) dataset (2.13GB)

link

Pretrained models (139.7MB each)

Enet

Tnet

DocUNet benchmark results

docunet_benchmark_paperedge.zip

The last row of adres.txt is the evaluation results. The values in the last 3 columns are AD, MS-SSIM, and LD.

Infer one image.

Download the pretrained model to the models directory.

Run the demo.py by the following code:

$ python demo.py --Enet_ckpt 'models/G_w_checkpoint_13820.pt' \
                 --Tnet_ckpt 'models/L_w_checkpoint_27640.pt' \
                 --img_path 'images/1.jpg' \
                 --out_dir 'output'

The final result: