0

Picture link I have to pre-process some documents for automated data extraction through pytesseract, but many of those images have text areas that are differently skewed.

Because of this most methods are not working which detect and de skew the image by calculating a uniform skew angle. The whole image has text areas that require de-skewing with different angles.

I am working with python and have tried the methods down below

  1. http://aishelf.org/deskew/

  2. https://becominghuman.ai/how-to-automatically-deskew-straighten-a-text-image-using-opencv-ama0c30aed83df

  3. https://pyimagesearch.com/2017/02/20/text-skew-correction-opencv-python/

Atish
  • 1
  • 1
  • 1
  • Are you working with scans or photographs of the documents? From your question I'm not sure if the issue is that you have multiple paragraphs of text under different angles, or you need to perform projective transformation because the text was photographed with a camera. – zteffi Jun 14 '22 at 09:15
  • 1
    just show us some pictures – Christoph Rackwitz Jun 14 '22 at 13:35
  • @ChristophRackwitz I have added a picture, in this one, the sheet is creased from the middle causing a different angle of slanting in different text fields in this image. – Atish Jun 14 '22 at 15:31

0 Answers0