1. Summary

I have third-party poor quality scanned PDF's, where start characters of the lines are highlighted.

I have numerous bad scanned PDF's, manually alignment of characters for replacements takes a lot of time for me. I can't find, how I can quickly overwrite/replace/repair bad quality characters.

2. Expected result

  • Before:


  • After:


3. Example data

It would be nice get expected result for any of these files.

Start characters corrupted:

Bad quality text

4. Limits

  1. I can't re-scan poor quality pages or make photos, because I don't have access to books.
  2. Please, any free Windows software. Not Photoshop, FineReader or Acrobat.

5. Examples of expected behavior

I have corrupted symbols; I don't think, that possible repair them via features as “Dodge/Burn”.

5.1. Text editors behavior

I remove previous text → I print new text as in any text editor:

Text editor

5.2. Align by line

I remove corrupt characters in image editor use “Bucket Fill” → I get font and font size, use pdffonts → I set them in image editor settings → I write the text:

Not aligned

→ I press to magic button “Align by line” → text is aligned between imaginary red lines:


I can't find feature similar as “Align by line” in any image editor (see “Image editors” section); I need align text manually, that take a lot of time.

6. Not helped

I can't find expected features in these programs. Also, I tried to use FineReader alternatives as NAPS2 and GImageEditor → no success.

6.1. PDF editors

I searched software here:

  • PDF-XChange Editor
  • LibreOffice Draw — I found, how I can edit OCR layer, not original PDF:

    LibreOffice Draw

  • PDFEdit — text editing doesn't work for me:


  • Icecream PDF Editor

6.2. Image editors

I searched software here:

  • GIMP
  • Inkscape
  • Paint.NET
  • Krita
  • 1
    I do not believe there exists a "quick" or "easy" automated solution for this.
    – Scott
    Aug 25, 2019 at 18:53

1 Answer 1


There's no software which can fix bad raster scans of text automatically. There's no easy way to align retyped text to lines of raster text, except by eye, or by using guides. Raster images are not editable text, so there's nothing to align typed text to, except pixels.

OCR software will not work properly if there are characters which are messed up. However, if you can just run it through some OCR software anyway, you could just retype the bits that got messed up.

I used this online service on one of your images. https://onlineocr.net/ . It has a Russian language option. Looks mostly OK, although it has errors because of the messed up characters, and those will need manual retyping. I have no affiliation with the site or its developers.

Here's the result:

АРИСТОТЕЛЬ , L (384-322 до н.э.) древнегреческий философ' и ученый

л 335 году до н.э. Аристотель, закончив воспитание с�оro Александра Македонского, основал в Афинах свою о� -хофскую школу. когда Александр неожиданно умер, в Афинах взяли верх *х:.нвники Македонии. Аристотеля, как прежде Анаксаго-л Сократа, обвинили в нечестии, то есть в неуважении s �•згач. Он оставил Афины, «чтобы афиняне, - как он ска-г= — вторично не совершили преступления против фило--:лии ,. казнив его, как некогда казнили Сократа. Философ гхелился на родине своей матери, в городе Халкида на :сове Эвбея. Здесь он и умер от болезни желудка. Много веков спустя появилась легенда о смерти фило-:_а из-за неудовлетворенного научного любопытства. Халкида расположена на берегу пролива Эврип, отделя-:• :его остров Эвбею от материковой Греции. Пролив этот -=- оычайно узок: его минимальная ширина всего 38 ме--+,в. Каждые б часов направление течения в проливе ме -ется. Аристотель долго пытался разгадать причину такой ь -оиаiии; когда же ему это не удалось, он от отчаяния бро-хтся в море, воскликнув: — Аристотель не смог охватить умом Эврип, так вгть же Эврип охватит Аристотеля! Эти слова привел византиец VIII века Косма Маюмский комментарии к «Стихотворениям Григория Богослова. 39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.