The Open library Project

[shachi]

Hello everyone!
The Open Library Project: http://openlibrary.org/

seems to have very neat features, but the feature that I am most interested is in the text-search feature(click on a book). As one can see it searches the book and highlights the text in the book, which is an image. How can such an effect be possible? Anyone have any idea how one can search an image for matching text? Just curious.

Thanks.

[Twey]

The search doesn't work for me so I can't be entirely sure, but I'd suspect the highlighted page images are generated server-side, and the text of the book is also stored.

[shachi]

The search doesn't work for me so I can't be entirely sure

Oh.

but I'd suspect the highlighted page images are generated server-side,

True, but how unless they've stored the text of the book as well.

I'd love to know how it's done without actually storing the text.

Thanks for your reply.

[Twey]

and the text of the book is also stored.

It is possible that it could have been done via a text-recognition algorithm.

[shachi]

Well, that seems pretty reasonable, I wonder how they do it. What I'd love to see is a text-recognition algorithm in PHP. That'd be quite a challenge.

[djr33]

I think it's stored serverside. That image may be rendered, as well. I doubt they are storing full images for every single page.
If you notice, the audio is just a recording of the book, so it's probably the same with text... just interrelated pieces.

[shachi]

drj33: True, I completely missed that.