Ideas for free: 100% correct OCR or an error correcting watermark

[snyders]

Present day OCR is good but still it makes 1-2% of errors. A simple solution would be to embed an error-correction code on each page as a small faint image, that would not distract from the main contents. This would allow to achieve 100% correct OCR of text.

One may ask, why bother if the text is prepared electronically in the first place. Counter-argument: in many cases we do not have access to electronic original, that's why we need OCR in the first place.

Update: turned out to be not a new idea, see comments.

Comments

[dimrub.livejournal.com]

A real life example is how the code of PGP made it (legally) out of the US.

[snyders.livejournal.com]

it probably was OCRed but was there error correction?

[dimrub.livejournal.com]

There was :)

[snyders.livejournal.com]

You are right. Thanks!