6

I am looking for an OCR software that does OK on older printed script text found in 1700s - 1800s texts and after recognition can store the original text and optionally translate it.

Currently I've been retyping printed books in German the best I can not being a German speaker and then once retyped using Google translate (which I know has its flaws) or manually translating them with a dictionary.

This is an example of the type of thing I am looking at from a Pennsylvania Newspaper in German from the 1790s. I am also looking at similar documents in German from what is now Germany or Switzerland in a similar printed script.

enter image description here

In this particular example above, and towards the end of when they stopped printing newspapers in German in the US they listed both English and German stories..

I am needing something that can provide some assistance to get a basic understanding of what something says in a semi-efficient manner to determine if a more precise translation is required and the examples are more likely to look something like this: enter image description here

This example is from Google Books of an old German text, but Google Book's OCR (as I can select the text and copy and paste it) of this type of text is very low quality to the point of not being useful. I also have resources not in Google Books.

Is anyone aware of a consumer OCR program, preferably for Mac, that has the capabilities to simplify this process for me?

  • Could you please post a picture showing sample script text? – Nicolas Raoul Jan 27 '15 at 7:58
  • @NicolasRaoul I added a few examples. – CRSouser Jan 27 '15 at 17:07
  • The Tesseract library can be trained for any font, so if you can find the same font, generate training images for it and train it you should be able to use that training data to hopefully get decent results. – user111 Feb 2 '15 at 8:02
3

There is a software named ABBYY and it's probably the software you're looking for. they even have software for Hand-writing recognition. I've been using this software since 2009 and it recognizes texts very good. there's another software which is one of the best in its level. Readiris 14 OCR Software that supports more than 130 Languages and it's for Cannon company. it also works on Mac but it does not translate the text for you. that's a whole different thing to be added to an OCR software. the software even recognizes complicated Asian languages.

2

Wolfram's Mathematica has built-in OCR capability.

It is available for Mac, but is free on a Raspberry Pi.

It does support the German language and it has functions to assist in decoding the image. I believe there are features within Wolfram Alpha to assist with translation although I've never used them.

And of course, there is a thriving Mathematica SE site to help out!

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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