Hide names in a scan

26 views
Skip to first unread message

Frédéric Monasse

unread,
Feb 13, 2018, 9:03:43 AM2/13/18
to tesseract-ocr
Hello,

I'm developing an app for lawyers who need to show official documents on their website but must hide the names of the people concerned.

The document is a scanned textpage in pdf format (image) with a lot of text inside where appear  the name of the concerned people example "M. Frederic Monasse". I need to hide these strings eigther blur or hide with a black layer (like in CIA secret files :-) )

I saw that tesseract was able to recognize text in images but is it possible to have the position of the text found in the image to put a layer at that position ?

Thank you for your answer.

Frederic.

ada...@turningcloud.com

unread,
Feb 15, 2018, 2:31:42 AM2/15/18
to tesseract-ocr
Hi Frederic

What you are looking for is possible, but tesseract might not serve your purpose. It is a topic of Natural Language Processing. Tesseract is an OCR so what it does is just reads out the text in an image. Tesseract dosen't know meaning to any word and if thats the case, how can someone read out the image that you will provide and then differentiate the names from other text. Hope you get what I mean.
You can look a way in NLP to solve your purpose.
Revert back for any queries.
Keep up the good work.


Regards
Adarsh
Reply all
Reply to author
Forward
0 new messages