OCR Scanners For Chess Books

Sort:
MGT88

I'm looking into OCR apps to pull PGN out of chess books; I'm able to pull text from pictures in OneNote 2013, however, the OCR accuracy is probably only around 80%, which isn't great for the purpose of PGN extraction/upload; does anyone have any recommendations in this regard?

mborko

Have you found something? I only found an app which scans the Board-Pictures, but not the PGN for the play, and the interesting variations to play.

Wondering if you were successful.

MGT88
mborko wrote:

Have you found something? I only found an app which scans the Board-Pictures, but not the PGN for the play, and the interesting variations to play.

Wondering if you were successful.

The quick/short answer is no.

First, I looked through many different generic OCR scanners; there are some pretty good ones out there, however, the better ones usually require a paid subscription, and there is still quite a bit of manual work involved in pulling text out of a chess book and getting it into a study.

Next, I looked through many different chess scanner apps and could not find any that featured text scanning.

I then came up with my own system/app/etc. idea (I wrote up specs detailing how the app would work etc.) and pitched it to a handful of existing chess app owners/developers/etc.; some agreed to look into building the feature, however, none did. I considered building the app myself (or paying someone to build it for me), but the costs are a little too high and it's probably not worth the investment.

Adding to all of the above; I have, until recently, always refused to use any digital chess books (I just prefer a physical book etc.); I have read over 50 chess books in paperback format over the last 2-3 years and I can say that the amount of labor/energy involved in either imagining the moves in my head, or in transferring the moves to a study, is immense. I recently decided to try an interactive digital chess book (meaning, moves/engine available in an analysis board while reading along) and quickly realized that I could read through it at 2x-4x the speed (maybe more) that I can through a paperback, so I will now use them wherever possible.

mborko

Thank you for answering my question. It's a pity, all those good old books, with great analysis only playable by hand or imagining it.

If you want, you can send me your specs and we can talk about implementing the ideas you've had. I would also introduce it to skilled IT students, they always need new project ideas. And some of them are also interested in chess, so maybe I can find a match.

MGT88
mborko wrote:

Thank you for answering my question. It's a pity, all those good old books, with great analysis only playable by hand or imagining it.

If you want, you can send me your specs and we can talk about implementing the ideas you've had. I would also introduce it to skilled IT students, they always need new project ideas. And some of them are also interested in chess, so maybe I can find a match.

Sure, send me your email address and I'll forward you some of the discussions I had with developers; we can also discuss the idea further.

fulgoren
MGT88 escribió: mborko wrote:

Thank you for answering my question. It's a pity, all those good old books, with great analysis only playable by hand or imagining it.

If you want, you can send me your specs and we can talk about implementing the ideas you've had. I would also introduce it to skilled IT students, they always need new project ideas. And some of them are also interested in chess, so maybe I can find a match.

Sure, send me your email address and I'll forward you some of the discussions I had with developers; we can also discuss the idea further.

fulgoren

HHello: ello:Hello:
I hope that the idea of digitizing books from pdf to pgn will be realized.
I am a blind chess player and I use a screen reader software that, through a synthetic voice, reproduces the text that appears on the screen.
These softwares do not recognise image pdfs. The best thing would be to have interactive books in pgn, with comments and diagrams in text accompanying the images.
So far the most accessible website is lichess, hopefully chess publishers will be interested in publishing digital books that include all of us.
This project is easy to realize today, since digital books were first textbooks hosted on computers (in pgn or chessbase formats). With physical books the work is more difficult, as it is difficult to minimise the errors inherent in any image-to-text conversion. The important thing would be to reduce the percentage of errors.
Thank you for this discussion.
Regards.
Carlos

Carlos