Did you try documents with variable layout?

wenwenyu / PICK-pytorch

Code for the paper "PICK: Processing Key Information Extraction from Documents using Improved Graph Learning-Convolutional Networks" (ICPR 2020)

https://arxiv.org/abs/2004.07464

MIT License

556 stars 193 forks source link

Did you try documents with variable layout? #22

Closed bravelywangexocr closed 4 years ago

bravelywangexocr commented 4 years ago

Hi wenwenyu,

Thanks this wonderful repo firstly.

One quick question, have you tried the documents/forms with variable layout? I just wonder if GCN can still help and perform well if the (relative) positional features are not consistent.

Thanks.

wenwenyu commented 4 years ago

Yes, We try it on the medical invoice having variable layout. Our experiments find that it can get the performance dropdown if we delete GCN compare with baseline method. GCN can help with extraction tasks even if the layout is not consistent.

bravelywangexocr commented 4 years ago

Get it, Thanks a lot!

wenwenyu commented 4 years ago

I am going to close this issue. Please feel free to reopen or create a new one if you have more questions.