Skip to content

Implementation of scanned document table segmentation with U-net

License

Notifications You must be signed in to change notification settings

dsabarinathan/DocumentTableSeg

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Resnet U-net based Scanned Document Table Segmentation

Overview

Data

ICDAR 2013 Table Competition you can download the train and test data from the below link.

https://roundtrippdf.com/en/downloads/

Pre-processing

The dataset contains 27 PDFs, first step is to convert the PDF into images. When we trained the model with original RGB images the segmented result is not good. Applied the distance transform on RGB images before passing into the model. It improved the overall segmentation accuracy.

Proposed method

alt text

About

Implementation of scanned document table segmentation with U-net

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages