---------- Forwarded message ----------
From: Jack Graham <friarjak@xxxxxxxxx>
Date: Fri, Jun 10, 2016 at 10:35 AM
Subject: Re: [vi-kindle] Newsprint images and OCR text reognition
To: William Brandes <williambrandes@xxxxxxxxx>, "vi-kindle@xxxxxxxxxxxxx" <
vi-kindle@xxxxxxxxxxxxx>
Cc: Gaby de Jongh <gabyd@xxxxxx>, Christopher Moss <
christopherhmoss@xxxxxxxxx>, Jack Graham <friarjak@xxxxxxxxx>
Hi all,
I believe I am the helper Kit mentioned. The https://www.newspapers.com/ site
provides a pdf page which contains an image of the microfilmed newspaper.
The Free OCR | Convert PDF to Text, Image to Text, PDF to Word, JPEG to
Word - Free Online OCR <http://www.ocrconvert.com/> system sees the pdf
frame and translates it, but only provides a place holder for the newspaper
image on the useless frame information OCR transformation. So the OCR is
translating the pdf perfectly but not the image embedded in the pdf. The
embedded image of the newspaper is what we want to be OCR translated.
I can not find a way to separate the image from the pdf frame and store it
as an image that can be translated by the OCR program. I did successfully
make a screen shot of the newpaper image and the OCR program did translate
it perfectly. But I can only shoot about a 10th of a newspaper column or
the print becomes too small for the OCR program to recognize. Thus for a
full column takes ten cycles of position the image on my screen, shoot it,
save it, feed it to the OCR program and translate it, Save it and repeat.
Onerous at best and too time consuming to sew the text back together and
send it to Kit so his text reader program can voice it to him. So far it
has been faster for me to read the articles to Dragon Natural Speaking,
edit the malapropisms Dragon generates, and send it to Kit.
I hope this clarifies our problem.
Jack
------------------------------
*From:* William Brandes <williambrandes@xxxxxxxxx>
*To:* vi-kindle@xxxxxxxxxxxxx
*Cc:* Gaby de Jongh <gabyd@xxxxxx>; Friar Jak <friarjak@xxxxxxxxx>
*Sent:* Friday, June 10, 2016 10:45 AM
*Subject:* Re: [vi-kindle] Newsprint images and OCR text reognition
hi chris. not sure i can help, but ...
is this newspaper being read as an online document?
what is meant by double image?
can your helper make sense of it?
how is this file/document referenced in your browser (if online)?
is the image embedded into a pdf file?
or, are the problems above you reference after using the online ocr program?
stay well ... william
On 6/10/16, Christopher Moss <christopherhmoss@xxxxxxxxx> wrote:
I am counting on you clever folks to help me figure this outgraphical
I am trying to read images of newsprint from the Times – Picayune
newspaperfrom the 1860s. I found it the free tool athttp://
www.ocrconvert.com/but it appears that it will not work with this site
since the image is a double image according to my helper.
Does anyone here have any tools or know of any tools for reading
text?this.
It is unfortunate but as an author I absolutely have to be able to do
I can't do my research unless I can read these newspapers.
