Re: Ways to scan and OCR notes taken in books (digital dump of "highlights") ?

 All the OCR software that has been mentioned  can do what you want to
a degree.   I have used Omnipage to scan technical articles out of 
mag.   The article may only take up part of a page.  You just select
the part of the page you need using the  softwares user interface.
But when you talk about highlighing with a marker and then have the
software detect the highlight. I am not aware any any software that
will do that.   But another problem with highlighting the text.  the
highlighted text may not scan in has clean has non highlighted text.
this could cause the OCR software to have errors when it  trys to
recognize the text.    Non of the OCR software that I have used,
converts to text with zero errors, 100 % of the time,  there is alway
someting that you will need to correct. 
 

On 27 Oct 2003 19:48:55 -0800, [EMAIL PROTECTED] (C_K) wrote:

>Thanks for the responses.  To clarify what I was asking about, I
>wondering about software / products / strategies that would do OCR
>_just_ on parts of a test that one is interested in, not the whole
>page.  Anyone know have any ideas there?
>
>A good ole pen highlighter would be the ultimate analogy -- _just_ the
>parts that one highlights gets dumped into a digital file.
>
>The pen scanner is the most obvious way to do this that I know of, but
>the Quicklink is _way_ too slow (I upgraded the pen's OS to current
>(2.43, I believe) and still too slow).  Anyone have good luck with
>other (hopefully not terribly expensive) scanner pens?
>
>What about OCR software that has the functionality to just do OCR on
>text that has been highlighted by an old fashioned highlighter pen? 
>Anyone know of anything like that?
>
>So, one would use a regular highlighter to highlight stuff from their
>book (nice and fast and portable and cheap) and then use a flatbed
>scanner to scan pages of the text that one did highlights on, and then
>the software extracts only the highlighted text.  Not ideal, but sure
>a hell of a lot better than having one's notes essentially hiding in
>the closed pages of various books....
>
>
>Thanks again 
>
>
>
>
>Anders Thulin <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>...
>> CSM1 wrote:
>> 
>> > when you get the hang of it, it is pretty easy. The big bad, is OCR is only
>> > about 99% accurate and only works on typed or printed text.
>> 
>>    Jim Weiler of Xoogi just announced what I take to be 99.999%
>> OCR work (perhaps I lost a '9') on the Bookpeople list -- he wrote:
>> 
>>    "Yesterday I proofread the first book I've done with only two mistakes in 
>> 200-plus pages."
>> 
>>    FineReader (unknown version) on 400dpi greyscale scans of a very well
>> printed and preserved book, it seems. No details about how long much work
>> it took in scanning, so it's difficult to say if it was done economically
>> or at very great expense in terms of time.