Physical Text To Digital Text

Question

Let me first prefix this question with, I have no idea which StackExchange site would be most appropriate for this question, but I thought LifeHacks might work...

In 1998 my Mom's side of the family had a big family reunion (my great grandmother had ten kids so it really was a big gathering). One of my distant Uncles wrote a sort of book on our families history for this reunion and my Mom gave me the book to read. I couldn't believe how big it is and how much research went into the book. I would like to figure out a way to get the whole book on a website I'm going to make where I can then share it with everyone in our family and ultimately preserve the history for longer than this flimsy book can.

In the hope that I don't have to type word for word this entire 300 page book is there a way where I can just scan the pages and get them in digital text? Obviously I could just take pictures and make the website using the pictures but I feel like it would be more beneficial to have it as actual text because then it might show up on Google searches better when someone searches a family members name or something like that. Also, if someone in the family ever does maybe a research project on our family they'd be able to copy some of the text and reference it more easily.

So does anyone know the best way I can get this old family book into digital text?

Answer 1

Use an android phone and the "Google Lens." feature recently added to the photos app and the camera photo review scrwwn.

OCR via Google Lens is pretty amazing and accurate beyond any OCR software I've ever used.

Below are some screenshots outlining the procedure using a cheap (100 USD) Nokia 3, best phone I've had the pleasure to use since my beloved Nexus 4 gave up the ghost.

I will detail a sample OCR scan of a Greek ethymologies book printed in 1976 that I dare not tear apart for scanning, that seems to have similar character density and typeface.

I took this original picture in less than ideal lighting conditions, using all auto settings on the so-so phone camera, no pecial photo techniques or fixtures to enhance the outcome were used, you could say it's just a plain amateurishly taken phone pic of a book's page. (Just make sure text is focused, no OCR will decipher blurry off-focus text)

Click on the Google lens icon, available via the preview after taking the picture or on the photo itself using Google photos app

Here's -Skynet- ^M^M^M^M^M^M I mean, Google Lens doing its magic scanning (the dots are a bit creepy but they had to do something to let you know that the googley AI is doing its thing, I guess)

Once the image is scanned, you'll find the text areas that Google Lens found on the picture clearly outlined and their text already extracted to the lower half of the screen. if you only want some areas and not others, just touch your selection to activate/deactivate them.

If you touch the extracted text, it will be placed in your clipboard for copy/paste goodness anywhere on your phone.

Afterwards, just paste the text on a Google docs document. There, you can: - correct any mistakes right there or on your PC, - share the document to your heart's content, - publish it as a web page with live updating of your edits,or - export to - plain text, - word document, - open office document, - kindle compatible epub electronic book with reflowing text, or - good ol' Non-DRMd PDF

It could be argued that this is probably the shortest route to publishing, with the widest possible output choices.

You could do it all from a single device, (Android phone with the appropriate apps installed) and be done with it in no time with a high accuracy rate, basically for free.

Here is the Google docs pasted Fragment

Here is the Google docs URL share, feel free to comment. You could also have someone help you edit the document remotely and simultaneously.

https://docs.google.com/document/d/1aizUDOHerSraU3fIw6lHLabmLSNsQ7PMXOl1IHHE0RU/edit?usp=drivesdk

Finally here is a Google Sites website published using the aforementioned document as linked source

https://sites.google.com/h-lo.me/ocrsample

It's https, desktop and mobile enabled and depending on tastes, generally not an eye sore. Not bad for 15 min total work and no coding whatsoever.

---

There is one refinement left, and that is to create proper paragraphs on the Google document, since Google Lens inserts a hard return after every line of extracted text, which makes every line it's own paragraph and this will become an issue if you want to use Google Docs features such as the table of contents, or when you export your document to a kindle compatible e-pub electronic book (messes up reflowing text)

You can just join every line where appropriate by hitting the backspace key on every line start, or this could be automated with a script.

So, I'm writing an apps script add-on that I will publish shortly to automate this process. I'll let you know here when it's done.

Answer 2

You can do this in stages. Start with putting everything online as page scans and updating as and when you can. The cerlox™ plastic comb binding makes taking it apart and putting it back into the binding easy.

As the printing appears to be normal serif type in the same size, the scans can be digitized by using Optical Character Recognition software. OCR can give you a draft text file that you can proofread and publish on the website for its final form.

At the same time, you can tidy-up the pictures and other pictorial material.

You can do this as time/resources becomes available to you for the project.

Answer 3

Some good answers here for approaching it yourself.

I would like to add my experience of paying someone else to do it for you.

I used Digitise My Books in the United Kingdom (I am UK based myself).

I was very very pleased with the results: Each book is returned as a PDF which has searchable (and copyable) text. A standard PDF technique is used whereby the original image for each page is retained but with a text overlay, such that you can highlight the original text on the page. Very good value. As someone overseas from UK, you can still send them the books.

They also offer an option for the book to be in editable word document format, at additional but very reasonable cost.

If you do not require the original to be returned, then the cheapest option would be to choose destructive scanning. This is where the pages are taken individually from the book and scanned. By default, the original book is not returned, though I believe you can request it, possibly at additional cost (e.g. for return postage) but the pages will be loose, having removed to be scanned individually. Destructive scanning is the option I chose for all my books and I did not require the originals to be returned.

They also offer non-destructive copying if you do require the original, but the cost is higher. They also accept your own digital scans if you have already scanned a book yourself - they can make this into the searchable, copy capable PDF or Word doc.

Have a look around their website. I really think that this is the best option: spend money to save time, rather than spending time to save money.

I don't work for Digitise My Books, nor have any financial interest in them (shareholder or otherwise).

Originally I had started to "scan" in the books myself, by photographing using a DSLR Camera (photographing is faster than flatbed scanning) with each page held open with a clipboard clip and blu-tak. But I found this to be rather labour intensive.

If you are still keen on doing it yourself, ScanTailor is an Open Source Windows Application that will format, split double pages/pairs of pages as scanned into individual pages, straighten and "dewarp" them. So that the resulting pages appear flat and straight as desired, however it does not do OCR: the results are still bitmap images. But at least it goes some way to batch automate tidying up any distortion of the pages, particularly of non-destructive copying where it is difficult for one to arrange the pages to be completely flat for large books.

Updated

Added more information about scanning options offered by a service. ScanTailor further info. Grammar corrections.

Answer 4

The fastest way to do this is to contact your relative and see if they still have the original files they used to create that book. From the front page photo I'd say it's been made on a computer. Convert from {insert really old word processor package here} to a current format and you're done.

The second fastest way of turning a stack of printed material into a digital document:

1. Remove the binding.
2. Cut off the left edge of the pages to get rid of the holes. Holes interfere with a document feeder.
3. Go through the book and unfold any creases and other damage that will interfere with a document feeder.
4. Find any reasonably modern duplex printer which has a document feeder and a scanning feature. Scan to PDF.

Then use any OCR package to turn the scanned pages into a Word file. I use the OCR features in the full version of Adobe Acrobat for this purpose, but there are many OCR engines around.

Answer 5

You might want to try a very inexpensive service: preserve-your-memories.info . When doing it myself, I use my scanner to scan to OmniPage, an OCR program, and then save as a pdf file which is completely searchable. Since your publication is plastic comb bound, it's easy to take apart to scan individual pages & then rebind. Taking pictures as outlined in the above suggestions is also very workable - a good one among many approaches.