Re: j Macarty, could use your assistance! was: Re: Introducing PDF2OCR and seeking testers

  • From: Kevin Jones <kevin@xxxxxxxxxxx>
  • To: programmingblind@xxxxxxxxxxxxx
  • Date: Sun, 16 Sep 2007 16:40:42 -0500

no ocr does not work well against capchas, blurry lines running through the font mess up any ocr program totally. 



Macarty, Jay {PBSG} wrote:

I agree with inthane; I suppose it is possible. Although, given that it
isn't usually hard to throw off most OCR packages with some unusual font
or background, I'd not have high expectations of getting the code right.


-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of inthaneelf
Sent: Saturday, September 15, 2007 8:14 PM
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: j Macarty, could use your assistance! was: Re: Introducing
PDF2OCR and seeking testers
its possible, stress the possible, I would try selecting the graphic, copying it and pasting it into an open paint file, so your not trying to dig the image out of the surrounding text that you already have access to, then run printscreen on it there.
and I say might, because some places, like Microsoft I know, use blurry backgrounds and oddly formed but readable (to someone with really good eyesight) forms of the numbers and letters, as well as sometimes putting something crossing the background to blur the letter formation, to keep them from being scanned and auto inputted by programs, which is the whole reason there doing that, to block automated generation bots. 

HTH,
inthane
. For Blind Programming assistance, Information, Useful Programs, and
Links to Jamal Mazrui's Text tutorial packages and Applications, visit me at: 
http://grabbag.alacorncomputer.com
. to be able to view a simple programming project in several programming

languages, visit the Fruit basket demo site at:
http://fruitbasketdemo.alacorncomputer.com
----- Original Message ----- From: "Dale Leavens" <dleavens@xxxxxxx> 
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Saturday, September 15, 2007 10:31 AM
Subject: Re: j Macarty, could use your assistance! was: Re: Introducing PDF2OCR and seeking testers 


So,

Is it likely this could be used to read those graphics used to secure
Web sites, the image of a password or number the user must enter to gain admission? 


Dale Leavens, Cochrane Ontario Canada
DLeavens@xxxxxxx
Skype DaleLeavens
Come and meet Aurora, Nakita and Nanook at our polar bear habitat.
----- Original Message ----- From: "Jamal Mazrui" <empower@xxxxxxxxx> 
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Saturday, September 15, 2007 12:28 PM
Subject: RE: j Macarty, could use your assistance! was: Re: Introducing PDF2OCR and seeking testers 



Thanks, Jay and Inthane.  I did some web research and found out the
following.

The current version is PrintKey Pro 1.04, costing about $20 and

apparently

published in 2003.  A 30-day, fully functioning demo is available from
http://www.warecentral.com

I installed the demo and converted the CHM help file to text in the
archive at
http://www.empowermentzone.com/pkey_doc.zip

Several graphics output types are possible, though Microsoft Document
Image format is not mentioned in the documentation or listed in the
options dialog.  TIFF is an output type, which I then tried to convert

to

text with Tesseract-OCR.  Unfortunately, the result was gibberish,

even

though Tesseract recognized it as a TIFF type it could handle.  I

suspect

the problem may be one of resolution since Tesseract reported only 96

MDI.

Unfortunately, however, I have not found a setting that can increase

TIFF

resolution.  If anyone else does, please let me know.  Although the
program would not be free, at a modest cost one could tie together
components that let one perform OCR on a window that a screen reader

fails

to incorporate readably in its off screen model.  The converted text

could

automatically be presented in a message box, multiline edit box, or

JAWS

user buffer.

Jamal

P.S.  I was confused by the one from Grab Bag partly because the
PrintScreen key was not bringing up the program for me, and there had

not

been any messages when I ran the program.  It turns out that my

computer

does not let PrintKey grab the PrintScreen key, but this is a known

issue

with some computers, so the program allows one to change the hot keys.

I

later learned about Alt+PrintScreen for the active window only, which

does

work in the current demo.  The program is also available in the system
tray.
On Sat, 15 Sep
2007, Macarty, Jay  {PBSG} wrote:


Date: Sat, 15 Sep 2007 07:48:50 -0500
From: "Macarty, Jay  {PBSG}" <Jay.Macarty@xxxxxxxx>
Reply-To: programmingblind@xxxxxxxxxxxxx
To: programmingblind@xxxxxxxxxxxxx
Subject: RE: j Macarty,
     could use your assistance! was: Re: Introducing PDF2OCR and

seeking

    testers

inthane,
You have summarized the features of PrintKey nicely; thanks. Just a
couple of additional notes.

1. The most recent version of PrintKey is a licensed software package

(I

don't recall the cost at this point). The latest free version was
version 4.0 which is the one I posted about originally.

2. My most common method for invoking PrintKey to capture a window is

to

launch the software by selecting the printkey.exe program in Windows
Explorer or FileDir.  Then, I normally use the keystroke to capture

only

the specific window with focus; the key is alt+printscreen. That way

I

only get the window I am interested in and not everything around it.

3. Once you press printscreen or alt+printscreen, you will hear a

sound

like a camera and then the PrintKey window will pop up. I then select
the print button and my printer dialog is set to go to a .MDI

document

instead of a regular printer. It should be noted that one other

useful

action when the PrintKey dialog comes up is to press ctrl+c to send

the

window image to the clipboard. Then, you can paste it into a document

or

an e-mail.

4. When the print selection creates the .MDI document, you will
automatically be placed in the Microsoft image viewer program. You

then

select the Tools menu and press enter on Send this image to a Word
document. You will then be asked if you want to perform OCR on the
image. I am not sure if you can override the OCR program the viewer
invokes at this point but it would be worth researching.

I'm thinking that, if PrintKey has the ability to save the image in
another format such as tif, it might be possible to try running the
image through the tool included in Jamal's PDF2OCR package to

compare.

My general impression is that the OCR included with Office Tools

isn't

the greatest. PrintKey gives you the ability to play with some

aspects

of the image in order to get it like you want and some of those

setting

may allow for better OCR translation. I just haven't been able to

find

what might improve it.



-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-bounce@xxxxxxxxxxxxx] On Behalf Of

inthaneelf

Sent: Saturday, September 15, 2007 1:58 AM
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: j Macarty, could use your assistance! was: Re:

Introducing

PDF2OCR and seeking testers

Jamal,

all I have at this time is the information that j Macarty's E mail

gave,

which I included in the text around the link. its a stand alone
application,
and if I remember right, one starts the application by entering on

it,

then
uses the screen print key on the image one is trying to OCR and I
believe
the application will open then to ask what you want to do with this
captured
image.  and I believe that you save it as a .MDI document, then you

open

that file which should start the correct Microsoft  program to OCR

the

file
and render it for your reading.

if j catches this post maybe he can clarify this whole thing further
than
this, and I'll nab his explanation and put it with the application

for

folks
to get with it.

I set it up and ran through it once, found that it seemed to be

working

alright, and haven't put it back on my newly reloaded machine yet due

to

the
fact that I rarely deal with screen images, the old "once in a

harvest

moon"
scenario.

there is stuff on the web about it though, I found the app in a

Google

search after J's post on the list about it and its value to him.

sorry I can't be of more help at this point, but if he doesn't have,

or

he
misses this post and doesn't expand on it, I will see what I can

find.

inthane
----- Original Message -----
From: "Jamal Mazrui" <empower@xxxxxxxxx>
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Friday, September 14, 2007 8:21 PM
Subject: Re: Introducing PDF2OCR and seeking testers



Thanks for the info.  Can you distribute more documentation with

it?

I

guess PrintKey.exe is the utility, itself, rather than an

installer.

When I ran it, there were no messages, either from a GUI or a

command

prompt.  I tried typical help command-line parameters to no avail.

What

is .mdi format (I have not heard of it before)?  Can any version of
Microsoft Word read it?  How does one do this with Word?

Assuming the program works as you have described (though I need to

learn

specific steps of invocation), it seems considerably different in

purpose

than PDF2OCR.  It seems that PrintKey is for learning about a

screen

image, whereas PDF2OCR is for learning about a PDF file.  Both

involve

OCR

(though PrintKey does not do this, itself), but the types of

information

to be accessed in a typically temporarl, static screen image are

usually

different than the content of a potentially large or formal piece

of

writing in a PDF.

Jamal
On Fri, 14 Sep 2007,
inthaneelf wrote:


Date: Fri, 14 Sep 2007 16:03:32 -0700
From: inthaneelf <inthaneelf@xxxxxxxxxxxxxx>
Reply-To: programmingblind@xxxxxxxxxxxxx
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: Introducing PDF2OCR and seeking testers

you can google for it, but I have put it up on my grab bag site

for

download, and the text that accompanies the link tells you how to

use

it

for
this purpose.

its a self contained executable, that you generally would go and

click on

to
use, I put it in a folder in my programs files folder and

generally

put a

shortcut to it in CO:\Documents and Settings\[my user named

folder]\Start

Menu\Programs\Startup
or for a universal (all users of the computer run on it:
CO:\Documents and Settings\All Users\Start Menu\Programs\Startup

this loads it at start up and makes it available by just hitting

the

old

"print screen" button on your keyboard.  but you can just make a

shortcut

to
it anywhere, and start it if you need it, so your not loading

unneeded

items
at boot up.

HTH,
Inthane
. For Blind Programming assistance, Information, Useful Programs,

and

Links
to Jamal Mazrui's Text tutorial packages and Applications, visit

me

at:

http://grabbag.alacorncomputer.com
. to be able to view a simple programming project in several

programming

languages, visit the Fruit basket demo site at:
http://fruitbasketdemo.alacorncomputer.com

----- Original Message -----
From: "Eileen Lafond" <Eileen.Lafond@xxxxxxxxxxx>
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Friday, September 14, 2007 2:38 PM
Subject: Re: Introducing PDF2OCR and seeking testers


Where do you get print screen?

Eileen La Fond
Phone (206) 386-0011
e.mail Eileen.LaFond@xxxxxxxxxxx


"inthaneelf" <inthaneelf@xxxxxxxxxxxxxx> 9/14/2007 2:29 PM >>>

I understand, print screen is a small utility that Jay Macarty ,

put

out

as
a interim solution for folks that had to deal with screen images

sent

to

them by clients, to OCR them, so I thought in the spirit of seeing

how

this
free OCR module that your using did compared to other scanners

that

are

free
also I would run it through it.

and yes, the setup I am talking about is for OCRing a completely

graphic

image and attempting to render it into a text version.

I wasn't trying to be offensive Jamal, but testing your program

against

some
others isn't a bad idea, to give you an idea of how well it, and

this

freebee OCR module does compared to other things out here, that

are

free,

or
included in suite's that a lot of us have?

inthane
. For Blind Programming assistance, Information, Useful Programs,

and

Links
to Jamal Mazrui's Text tutorial packages and Applications, visit

me

at:

http://grabbag.alacorncomputer.com
. to be able to view a simple programming project in several

programming

languages, visit the Fruit basket demo site at:
http://fruitbasketdemo.alacorncomputer.com

----- Original Message -----
From: "Jamal Mazrui" <empower@xxxxxxxxx>
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Friday, September 14, 2007 5:29 AM
Subject: Re: Introducing PDF2OCR and seeking testers



FYI -- the 1.0 release version is considerably better than the

beta

version due to performing OCR on a page-by-page basis and to the
convenience of converting any number of PDFs in a directory with

a

single command.  PDF2OCR does work on text-based as well as

image-based

PDFs, but the results on text-based ones are generally not as

good

as

with other utilities that extract the text directly rather than
analyzing the picture of each page.

Jamal
On Thu, 13 Sep 2007, inthaneelf wrote:


Date: Thu, 13 Sep 2007 15:52:12 -0700
From: inthaneelf <inthaneelf@xxxxxxxxxxxxxx>
Reply-To: programmingblind@xxxxxxxxxxxxx
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: Introducing PDF2OCR and seeking testers

and print screen, lets see how this freebee works out, smile

and Jamal, I appreciate what your working on, just want to make

sure

your
getting at least 60% of what your hoping for out of it.

inthane
. For Blind Programming assistance, Information, Useful

Programs,

and

Links
to Jamal Mazrui's Text tutorial packages and Applications,

visit

me

at:
http://grabbag.alacorncomputer.com
. to be able to view a simple programming project in several
programming
languages, visit the Fruit basket demo site at:
http://fruitbasketdemo.alacorncomputer.com

----- Original Message -----
From: "Lloyd Rasmussen" <lras@xxxxxxxxxxx>
To: <programmingblind@xxxxxxxxxxxxx>
Sent: Thursday, September 13, 2007 3:47 AM
Subject: RE: Introducing PDF2OCR and seeking testers



If someone wants to submit a sample, I can run it through the

PDF

tools
that
come with OmniPage 15.  Also, someone else should run the

sample

through
the
image printer and OCR built into MS Word.


Lloyd Rasmussen, Kensington, Maryland
Home:  http://lras.home.sprynet.com
Work:  http://www.loc.gov/nls


-----Original Message-----
From: programmingblind-bounce@xxxxxxxxxxxxx
[mailto:programmingblind-
bounce@xxxxxxxxxxxxx] On Behalf Of Jamal Mazrui
Sent: Thursday, September 13, 2007 5:42 AM
To: programmingblind@xxxxxxxxxxxxx
Subject: Re: Introducing PDF2OCR and seeking testers

I would expect better results from Kurzweil 1000, which uses
arguably
the highest grade commercial OCR technology.  The free Adobe

Reader

does
not do OCR, so would not make the sample file accessible.

Jamal

__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind



__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind


__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind


__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind

__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind

__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind


__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind



__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind
__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind


__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind



__________
View the list's information and change your settings at
//www.freelists.org/list/programmingblind

__________
View the list's information and change your settings at //www.freelists.org/list/programmingblind 
__________
View the list's information and change your settings at //www.freelists.org/list/programmingblind 


__________
View the list's information and change your settings at //www.freelists.org/list/programmingblind

Other related posts:

  1. Home
  2. programmingblind
  3. Archive
  4. 09-2007