This fork has been modified to work with:
- python 3
- updates to opencv
- used for vertical line detection instead of horizontal text detection
Page dewarping and thresholding using a "cubic sheet" model - see full writeup at https://mzucker.github.io/2016/08/15/page-dewarping.html
Requirements:
- scipy
- OpenCV 3.0 or greater
- Image module from PIL or Pillow
Usage:
page_dewarp.py IMAGE1 [IMAGE2 ...]