Charachter recognition .... How to ?

alun4

New computer, new software!

I would like to scan text in French and english from my HP Photosmart C4580 to my PC or laptop. Both are using Windows 7 (PC has home premium and laptop has Professional). I have Microsoft Office 2010 Professional Plus).

Can anyone suggest a step by step (idiot's hand hold!), please ?

fenlander_uk

You need OCR (Optical Character Recognition) software if you want to scan documents into editable text. Your scanner may have been supplied with suitable software (doubtful, though). If not, an investment will be needed as OCR is one area where the free options are really pretty poor. One of the best is ABBYY Finereader, but it will set you back about £90.

If you don't need the text to be editable, you can scan pages to pdf format. Essentially, these are graphics files - you can read and print them using Adobe Reader or similar, but you can't load them into a word processor for editing.

So the first question is - do you want editable documents?

alun4

fenlander_uk wrote: »

You need OCR (Optical Character
So the first question is - do you want editable documents?

Hi, Yes I want to be able to edit (and translate when it's in French).
Thanks, Alun

fenlander_uk

That's pretty much your answer, then. There's a free trial available at ABBYY.com. In very general terms, you open the Finereader program and use it to scan your pages. It recognises the text as it scans. When you've finished scanning, you can save the results in a variety of formats or drop the whole document into Word.

There is a free OCR package called FreeOCR (surprise). I've tried it and found it significantly inaccurate with any but the simplest text. OCR must be accurate to be useful, otherwise you spend more time correcting than you would if you just retyped the document. Note that you will always have to do some proof reading of OCR documents, especially if their accuracy is critical.

One other point: Finereader recreates the format of a scanned document pretty well, but in doing so it creates its own hierarchy of styles. These are not always the easiest things to manipulate afterwards.

Translation is another matter. If you're looking for machine-translation there are a number of online options available, but while they produce readable results (usually), they do not produce fluent or idiomatic language. Good (literary/business quality) translations are still a largely manual process.

alun4

If I get a "readable" document to word, I can send it by google email and then use their translate to give a "reasonable" indication of the "sense" of the doc .... with my night school French class to assist !

It would not do for technical/legal documents.

I am surprised that office 2010 has not got OCR.

buglawton

The full business version of Office 2003 had a good ocr add-in.
(Programs > Microsoft Office > Microsoft Office Tools > Microsoft Office Document Scanning)

This was a very accurate program compared to software given away free with early gen scanners.

And there are some web services that accept a file and send back the text, e.g.
http://www.onlineocr.net/
http://www.newocr.com/

just Google: ocr online

alun4

buglawton wrote: »

The full business version of Office 2003 had a good ocr add-in.
(Programs > Microsoft Office > Microsoft Office Tools > Microsoft Office Document Scanning)

e

Thanks, doesn't it seem strange that 2010 does not do what 2003 does! My first printer/scanner/copier/fax was a HP office jet pro and the supplied software had OCR capability. The current HP scans to PDF or TIF.