Tech: OCR (Optical Character Recognition)

Subject: Tech: OCR (Optical Character Recognition)
From: wysiwyg
Date: 30 Jan 10 - 12:37 PM

If this has been covered before, at Mudcat, please feel free to redirect me.

I used to have a scanner that would scan in a page of text, and offer the option of outputting either a .JPG or a .TXT file. It appears no longer to do OCR.

Other inexpensive software options for Windows XP? What I want is editable text.

~Susan

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 01:09 PM

Susan... http://www.simpleocr.com/

This is the usual freeware 'standard' offered. I 'may' have another which is better, but I haven't looked at it in ages. Let me poke about in the catacombs and see what the situation is. (I know there is a paid mode, but they gave one version away free awhile back)

Now, here is another which is listed and claims to be as accurate as any. I can't comment on it.

http://code.google.com/p/tesseract-ocr/

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 01:18 PM

Here is a page which compares several of the programs, including TopOCR, which require a 3+ megapixel digital camera, and works from that image.

http://www.makeuseof.com/tag/top-5-free-ocr-software-tools-to-convert-your-images-into-text-nb/

Subject: RE: Tech: OCR (Optical Character Recognition)
From: JohnInKansas
Date: 30 Jan 10 - 01:33 PM

In order to do an OCR on the scan, the scanner software would have needed an OCR program. If it did a conversion, the scanner driver program would need to know what OCR program to send the scan to, and where to find it. While the OCR feature might have been imbedded in the scanner driver, it's more likely that it was a separate program used by the scanner.

If you have program disks that came with the printer, you may be able to "reinstall the scanner," being careful to include the OCR if it's an optional feature.

If OCR was a "standard feature" for the scanner, you might be able to download an updated driver/program set from the scanner maker.

If you can't recover OCR capability by reinstalling what you have, you're best option might be an update to a newer version of the OCR program you had, since it's fairly likely to offer you a discount. You may be able to download "ReadMe" and/or manuals from the scanner maker that will identify the OCR that came with your scanner(?).

I have had a couple of self-standing OCR programs for OCR. With the older ones, you had to save the .jpg and then separately open the OCR and select the ones you wanted to include in the conversion. I usually saved the results as Word .doc files, since I found editing to correct the missed chars easier than in plain text.

Installation of the software for my latest "super multipurpose" printer/scanner installed IRIS OCR which is excellent, but the installation didn't inform me that it was also removing all prior OCR programs. My Nuance-Omnipage OCR program simply disappeared. It's possible, given this rather predatory behavior, that some other device you've installed has OCR capability and has removed the OCR program your scanner used(?).

The installed OCR that came with the new machine didn't allow me to open the OCR program separately and input exisitng .jpg files, but it did have significantly better accuracy than the ones it removed. I eventually bought an IRIS upgrade to get both the "OCR from scanner" and separate ability for "OCR from saved .jpg files."

The original IRIS software that installed with the MP programs allowed direct "save as PDF" and the text in scans was processed by OCR so that the pdf files included (mostly) editable text. Even the few characters this converter didn't recognize are embeded as glyphs in the pdf, so you only find them by highlighting when you select to copy to a wp program. I could have just printed separate .jpg (or any other) files and then scanned the prints to get conversion; but the number of files I have that might someday need conversion suggested that having separte .jpg file import would be handy.

The PDF files that my MP produces aren't editable with software I have, but the text can be copied and pasted to Word to make a new (editable) document.

The "big names" in OCR probably are IRIS, OmniPage, and Text Bridge. I have had all three. My current IRIS is the most accurate, but I can't be sure that isn't just because it's a much later version that reflects improvements all of them may have now. None of these are really "cheap," although you probably can get "starter versions" of any of them for around $100.

I'm sure there are freeware/shareware programs available; but I haven't looked at any of them. Getting your scanner to automatically send scans to/through a "foreign" OCR program may be something you'll have to work out - if it's possible with your scanner.

John

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 01:36 PM

Now... I have version 4.0b1 of This product, ABBYY Finereader. It was given away free somewhere.

There are now up to version 10, and the full version is expen$ive. I 'think' that the giveaway was a "download and get registration key within 24 hours" sort of deal. It does give one an incentive to see if the free programs will work.

(ABBYY has a 15 day trial period, if that would help)

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 02:16 PM

Yep... ABBYY 4 was given away a couple years ago here.

(Now & then they have something you really want. I got the full WinRAR there at one time.....but you do have to watch it regularly, as it is always a 24 hr. deal)

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bonzo3legs
Date: 30 Jan 10 - 02:22 PM

This trick works for MS Office and later - scan the text and save as a tiff file. Then open up the file which should open in Microsoft Office Document Imaging. Click Tools and xxxxWord. It nudges you to click the OCR button and scan appears as editable text in Word.

I am an accountant and clients often send me prints of spreadsheets. If printed with gridlines, follow the above procedure, then copy and paste into Excel for a fully functional spreadsheet - no formulas of course but a great time saver for me. If it has no gridlines the procedure will not work, so they need to be drawn in first.

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bonzo3legs
Date: 30 Jan 10 - 02:23 PM

In the above post it should read - This trick only works for MS Office 2003 and later.

Subject: RE: Tech: OCR (Optical Character Recognition)
From: wysiwyg
Date: 30 Jan 10 - 05:08 PM

OK...... (?!?!?!?)

The scanner is: HP5300C. As supplied it was said not be to compatible with XP. I got it bundled with the Gateway (Win ME), and later the puder was upgraded to XP (incl all 3 svc packs). At that point it also got a new HD, so all the old software was erased. (At that point I was not using the scanner.)

OK so the other day I got the scanner plugged in and it gave me two messages: not compatible with XP, and not recognized as installed. I had the old disc and installed it. Still not reckanized. Then I went to HP and downloaded an upgrade I thought would do the trick. Maybe I downloaded the wrong item-- they had many and I was having vision issues with fine print.

(We also tried swapping USB cords in case that was the problem but got same message: "can't find.")


Does this info help anyone give me VERY simple instrux on next steps?

Remember-- strokebrain plus PTSD plus vision issues.


Thanks!

~Susan

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 05:24 PM

hmmm.. It might just flatly not BE compatible with XP..I'll see what I can find out.

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 05:35 PM

Ok...it says driver 3.04 'should' make it work, but I see several posts by those who have had troubles. The first couple of explanations I tried to read were above me. Still looking....

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 05:42 PM

best info I can find on it (from an HP help site)

Subject: RE: Tech: OCR (Optical Character Recognition)
From: katlaughing
Date: 30 Jan 10 - 06:14 PM

Bill, thanks for the simpleocr link. It just solved a problem for me.:-)

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 30 Jan 10 - 06:20 PM

well! Always happy when something does some good for someone!

Subject: RE: Tech: OCR (Optical Character Recognition)
From: wysiwyg
Date: 31 Jan 10 - 11:11 AM

So it looks, from the HP site link, like it will just be easier to download new (free) OCR software?

~Susan

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 31 Jan 10 - 11:49 AM

As long as you can get a scanned image of the text you want, new software ought to be able to take it from there.

It looks like major hassles making the built in software output a text file.

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Mick Pearce (MCP)
Date: 31 Jan 10 - 01:36 PM

I have the 5300C on my desktop and it runs fine (and has done since I first loaded XP years ago) under XP/Pro.

The HP Page BillD linked above has the versions of the software I have installed (I've still got the downloads on my disc - sj664en.exe - that's the large driver and precision scan software file from the top of the page, sj669en.exe - that's the usb patch, since that's what I'm using;use the parallel version if that's how you've got it connected and sj700en.exe - that's the software destinations path from the bottom of the page. I installed these when I updated my system from Win98SE to XP - along with other drivers that the XP installation checkup told me would be problems after the upgrade).

So if you download them and run them the scanner should be fine.

Mick

Subject: RE: Tech: OCR (Optical Character Recognition)
From: wysiwyg
Date: 31 Jan 10 - 08:05 PM

Apparently the first time I downloaded the update software, it did not all download totally, since I could not find files labeled like MP described. So I did what MP said and when the huge file downloaded, I ran it, and BANGO! Something that works almost as well as I remember it.

Thanks, gang.

~Susan

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Bill D
Date: 31 Jan 10 - 09:03 PM

awww...wonderful! (See...we needed an expert, and Mudcat has several!)

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Joe Offer
Date: 06 Jan 12 - 09:17 PM

There's a nifty program called Microsoft Office Document Imaging that comes with every iteration of Microsoft Office. It works very simply and effectively. It's not part of the standard installation, so you have to go to "install programs," select "Microsoft Office," and "tools."
You can scan from it or open images and OCR them. Then you just highlight the text you want and copy it, and then paste it into a word processor. I've been using it for almost ten years, and I love it.
-Joe-

Subject: RE: Tech: OCR (Optical Character Recognition)
From: GUEST,999
Date: 06 Jan 12 - 10:44 PM

That's Giok with the sheep, not me. I don't care what the picture says.

Subject: RE: Tech: OCR (Optical Character Recognition)
From: Mick Pearce (MCP)
Date: 07 Jan 12 - 07:41 AM

999 - !!??

Mick