Image preprocessing library

Question

Image preprocessing library

zamazan4ik opened this issue 7 years ago · comments

Hello,
I am fan of your program, thank you.

I am working on image preprocessing before OCR. I see that you have a lot of stuff already implemented in textfairy. But i think that if will be available library with preprocessing, you will can concentrate work on other stuffs.

Are you interested in this library? I am implementing now it, based on opencv and my algorithms.

renard314 · Answer 1 · Fri Oct 20 2017 21:11:49 GMT+0800 (China Standard Time)

This sounds like a neat idea and if you open source your project I will certainly check it out. I have a lot of test documents in this repo that can used to compare performance.
One thing to look out for however is apk size.

Alexander Zaitsev · Answer 2 · Fri Oct 20 2017 21:29:19 GMT+0800 (China Standard Time)

https://github.com/ZaMaZaN4iK/PRLib

logos · Answer 3 · Sun Oct 29 2017 18:47:38 GMT+0800 (China Standard Time)

Maybe OpenNoteScanner can be of help too in your library. It uses OpenCV and is opensource as well.

renard314 · Answer 4 · Fri Feb 23 2018 20:40:14 GMT+0800 (China Standard Time)

@zamazan4ik your link is dead. If you are still interested I can setup my test routine (200+ images) to run with your preprocessing library. That way I can compare against the current baseline.

owied · Answer 5 · Sat Apr 27 2019 21:55:36 GMT+0800 (China Standard Time)

Do you mean: https://github.com/leha-bot/PRLib ?

renard314 · Answer 6 · Sun Oct 18 2020 04:53:22 GMT+0800 (China Standard Time)

I check the library and while it has useful methods it is not a replacement for the image processing pipeline that I have right now. What I need is a function that takes an image and converts it into a clean binary taking into account varying lightning, different text sizes etc.