Book scanning

[kusmi]

After receiving my iLiad I'm wondering how to get my content into the device: Most books I'm interested in, do not have an ebook version... So I'm thinking about scanning some of my books at home.

Reading some comments in this forum or on the internet, there are a few techniques out there, but none is really "kicking"

My conditions are:

• Non destructive (e.g. I don't want to cut my books in pieces)
• I'm happy with scanned images, so no ORC required
• I don't need high resolution, just be readable on the iLiad

I found several ideas how to do it (beside the really expensive methods which costs you a fortune):

• Using the OpticBook 3600 scanner (still somewhat expensive)
• Using a hand-scanner, e.g. RC800 but I guess this is quite fuzzy?
• Using a normal flat-bed scanner (big books will have problems, because of the book-fold)
• Using my glass table to put the book on and my digital camera underneath

The most cost-sensitive way looks like flat-bed scanner or digi-cam. But of course the produced images need to be post-processed - rotating the slighly misaligned pages, correct the distorted book-fold (see link of OpticBook what I mean here )

Does anybody know of software to help there? So far I have found Unpaper but it does not correct the distorted book-fold...

Any ideas?

[ath]

Originally Posted by kusmi

Does anybody know of software to help there? So far I have found Unpaper but it does not correct the distorted book-fold...

ABBYY FineReader Pro splits page spreads, rotates and deskews images ... the intention is to improve OCR, but as it can output image PDF files without any OCR, it can be used just for 'page straightening'.

It's probably overkill if you don't plan to do any OCR at all, though.

The simplest is probably to photograph single pages (flat pages - not page spreads) with a digital camera, and then add whatever postprocessing is required.

[kusmi]

Many thanks, so you mean "ABBYY FineReader" is also capable of doing a conversion

from:


into:

[yokos]

Yes, I should work; it worked for me some days ago.
A con of this method is: it takes ages, FR takes a lot of time to calculate.

Using a digital cam & a heavy glass plate for analogue minimizing page curves works better for me.

[kusmi]

Are there also other tools as FineReader?

[ath]

Originally Posted by kusmi

Many thanks, so you mean "ABBYY FineReader" is also capable of doing ...

No ... no program will do anyting like that automagically. If you start with clear scans that are just curved (typically from overhead scanners), or skewed, it will straighten them out. Of course, that kind of processing eats resolution, so the resulting image will be fuzzier, and the more work that is needed, the fuzzier the result.

Same thing with splits -- there should be at least some white in gutter. I'd expect FR to split that page into three columns: left, wide middle, and right, since that's where there is visible white.

Don't expect magic. If you want clear, straight images, you better begin by trying to get them that way from the start.

[Henry Loenwind]

Don't expect the RC800 to work for this. It is a fine device, scanning is fast, result is very nice, but: The image compression takes up to 1 minute (for colored A4) and it switches itself off after each scan. So you'd scan one book page in about 1 second, wait 10 seconds until it's compressed and switch it on (takes 2 seconds) again...

[kusmi]

I saw some research papers

Book Fold Correction

Distortion Reduction

Distortion Reduction 2

which should do that "automagically" and also some high-end book scanners:

High-End Book Scanner

If they can do it, then perhaps also some clever mind could do something similar

[Paul Moews]

I have an optiscan 3600 - It has some advantages - some helpful buttons -
scanning fairly near the edge - one disadvantage is that images are not
deskewed - also some books have print so near the crease that they are
difficult to scan a page at a time.

I find any method that does a page or two at a time very tedious - hours to
do a single book. It's best not to have to do a page at a time and you might
find as I do that it is important to deskew.

Perhaps you could buy two books one to scan and one to read. Alternatively
I find it quite easy to read a book that has been unbound for scanning. Keep
the pages in order inside the cover and close with a rubber band.

[Moonraker]

I have an OpticBook 3600 and an Epson Perfection 1670 which I use with Abbyy Fine Reader 8.

They are both fairly fast flat-bed scanners.

The Opticbook is only really suitable for hardbacks because it requires a 6mm gutter and most paperbacks have less than this. But it's great for large books scanning one page at a time.

I use the Epson for paperbacks, scanning two pages at a time and AFR is clever enough to recognise these as separate pages - so long as you press firmly on the spine of the book to ensure that both pages are as flat to the glass as possible.

Scanning at greyscale, 300 dpi is perfectly adequate for novels that contain text only.

I can scan a 200-page paperback in about an hour - but the time-consuming part is the proof-reading of it.

I have never scanned to PDF but I will try this some day. At least it would remove the proof-reading part.

[kusmi]

There might be a new low-cost tool out there to scan books - using your digital camera:

http://www.atiz.com/snapter.html

Let's try, when it is available (March 7) :-)

[RWood]

Great, they finally put a date on it. Before it was just a non-working "download here" link.

[slayda]

Originally Posted by Moonraker

Scanning at greyscale, 300 dpi is perfectly adequate for novels that contain text only.

I can scan a 200-page paperback in about an hour - but the time-consuming part is the proof-reading of it.

Moonraker,
I've found that a quick proof of the scanned pages in FR and then a once through in MS Word for grammar and spelling works for the worst errors. The I read it on my Palm-TX as the final proof. This way I can catch the small errors, mark them on the Palm and go back to the Word file to correct the master copy.
Doing it this way, I can enjoy reading the book and also get a good proff reading done.

[Bob Russell]

There's something similar and useful at http://www.scanr.com/
but that one does things on their server.
Also good for business cards, btw.

Is snapter local or do you have to upload/email your pictures?

[NatCh]

I think it must be local -- there's a (not yet functional) 'Download Snapter Trial' link on the More Info page.


I took pix of a short booklet (~60 pages out of a ~150 pages in the booklet) for my wife when she had to return it to the library before she was finished with the source (it was really old, so photocopying wasn't a good option). That was a real pain to do. I'm not sure how well I'd do with snapping an entire novel ....

I s'pose I could try the free trial on those pix and see what it gives me. Assuming she still has them, of course.

It should be noted that this does not appear to include OCR software:

(from the 'uses' suggestions on the 'More Info' page)

With OCR software, you can turn all of the books that you have into your own huge personal digital library. And now whenever you have an open book exam or are thinking of some words you don’t know just “search” for that with any desktop search program.