[Bell Historians] OCR (Was: 1st 10-Bell Peal)

[John Harrison john@jaharrison.me.uk [bellhistorians]]

In article <182492EB51E64D4E97E95C70D75B9099 at TonyVostroPC>,
   'Tony Smith' smith_a_p at btinternet.com [bellhistorians]
<bellhistorians at yahoogroups.com> wrote:

>  For searching to be possible  ,,,  PDF supports the overlaying of the
>  images on the OCR'd text

That is valuable because you have the feel of the source image with the
illusion of being able to search it and copy it.  But like all valuable
illusions there is a catch.  OCR isn't perfect and the imperfections are
hidden behind the authentic image.  If you copy text with errors you can
see them and correct them.  The problem is with searching.  We tend to
trust searches, but in imperfect OCR text will miss any corrupted word.   

OCR is much better than it was, as Tony notes.  I recently did a crude
comparison between the three Ringing World DVDs.  I picked a year from
each (1920 1950 and 1980) and then copied the text from the first page.
The first is quite bad.  The second looks OK superficially but a lot of
words have breaks in them.  So for example a search won't find
'fundamental' but it will find 'funda' or 'mental'.  The third was pretty
good but erred the other way with quite a few words concatenated. 

Working with the early discs I developed the strategy of searching on
partial words, but it's a balancing act because then you have more hits to
scan by eye to see which are what you were looking for.

-- 
John Harrison
Website http://jaharrison.me.uk


------------------------------------
Posted by: John Harrison <john at jaharrison.me.uk>
------------------------------------


------------------------------------

Yahoo Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/bellhistorians/

<*> Your email settings:
    Individual Email | Traditional

<*> To change settings online go to:
    http://groups.yahoo.com/group/bellhistorians/join
    (Yahoo! ID required)

<*> To change settings via email:
    bellhistorians-digest at yahoogroups.com 
    bellhistorians-fullfeatured at yahoogroups.com

<*> To unsubscribe from this group, send an email to:
    bellhistorians-unsubscribe at yahoogroups.com

<*> Your use of Yahoo Groups is subject to:
    https://info.yahoo.com/legal/us/yahoo/utos/terms/