ocr for screenshots?

[jim.s.witherspoon]

I was trying to extract text from a MozillaWindowClass window, and had no
luck with Nirsoft SysExporter, or with CopyMessageBox.

I got great results with JOCR - it doesn't extract the text but it captures
an image of the text and OCRs it.
http://home.megapass.co.kr/~woosjung/Product_JOCR.html
But JOCR requires MS Office 2003 or higher to be installed.

Is there another freeware program that gets such good results with
screenshots, but doesn't require Office to be installed? I tried SimpleOCR
and FreeOCR but the results were miserable. I'm sure that these programs
are expecting the line spacing of many more pixels (300 dpi) where as these
screenshots are around 100 dpi.

[rich]

Try TopOCR
http://www.topocr.com/topocr.html

Seems to have improved since I last tried it, OCR'd a screen shot - a
google search listing - reasonably well.

--
rich

[Daniel Menzel]

jim.s.witherspoon schrieb:

The thread tells it all. It´s not depending if it´s a screenshot, a
scanned document or anything else. OCR just scans for matches with fonts
at all. The higher the quality of the used picture data and ocr software
the better is the result IMHO.
I´m drunken at the moment, but I think it is that way.
Office is needed for the output format and for comparing with existing
known fonts installed on the OS.
Enhanced OCR software uses another and more improved algorithm that is
able to convert handwritten texts at all.
Even if you have a "Sauklaue" <- thats german for a very bad font
character.
At least, perhaps too much beer...
Best wishes

Daniel

--

-----------------------------
// It can be fun to use a pc \\
// :) :) :) :) :) :) :) :) :) \\
**********************************
* Sorry for my bad english guess *
* I will never learn to speak and*
* write that language correctly! *
**********************************
* www.efmenzel.de *
**********************************

[jim.s.witherspoon]

> Try TopOCR
> http://www.topocr.com/topocr.html
>
> Seems to have improved since I last tried it, OCR'd a screen shot - a
> google search listing - reasonably well.

Thanks Rich. TopOCR did much better OCRing my screenshot than SimpleOCR or
FreeOCR - not as well as JOCR, but then TopOCR does not require MS Office
like JOCR does.

jim

[Richard Steinfeld]

jim.s.witherspoon wrote:

>> I got great results with JOCR - it doesn't extract the text but it
>> captures an image of the text and OCRs it.
>> http://home.megapass.co.kr/~woosjung/Product_JOCR.html But JOCR requires
>> MS Office 2003 or higher to be installed.

I just downloaded JOCR. I was startled to discover that what I
thought would be an installer seemed to be, in fact, the program
itself: it ran. And it was amazingly tiny.

If I've got this right, it's a very efficient application, frugal
with resources. Is this correct?

Note, again, that JOCR is freeware but it runs in conjunction
with a MS Office module, so it interfaces with a common payware
app. It won't help people who use freeware alternatives to MS
Office apps.

I discovered a few years ago that with OCR applications, ability
to work with small fonts is where the rubber meets the road. A
free Xerox application that came with my scanner turned out to be
unusable garbage because it only read large faces. So, it was
"teaseware" for a high-priced upgrade. I discovered this after
I'd printed out its 1,000 page manual. And into the garbage went
two reams of paper.

Richard

[Johnw]

jim.s.witherspoon expressed precisely :

> Thanks Rich. TopOCR did much better OCRing my screenshot than SimpleOCR or
> FreeOCR - not as well as JOCR, but then TopOCR does not require MS Office
> like JOCR does.

Free OCR online scan
http://www.free-ocr.com/
FAQ
http://www.free-ocr.com/faq.html