Google opened the API for OCR

Yes, they work in the Google Books surely bring good results, we can begin to reap. And now I'll tell you how.

Scanned documents do not always require a transformation in the actual text. But sometimes she wants to not gain any agreement again, and an electronic copy of something and not. Of course, you could use some cheap program the OCR, going with the scanner, or even buy it (not steal you) FineReader. But the free desktop OCR software runs also because now the texts are increasingly being photographed, not scanned.

To download any pictures (jpeg, png, gif) with text recognition for subsequent Google API and opened in Google Docs. Now you can upload images to a document library, and the Google server will transform it into text.

There is also an application example that illustrates how the API:

But you, the programmers should think about creating their own interface to these capabilities. For example - do you have the scans of books in PNG? This is a perversion, is not it? So there you have the card in hand - write an application that loads the text page by page, does not violate the limits and connects the entire load in a single text.

But remember, there are limitations in the API, and the main of them, it seems to me - something that is recognized only Latin as soon as it. Also care must be taken to character height is not less than 10 pixels, and the total size of the image does not exceed 10 megapixels.