Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Computer Support > OCR program needed

Reply
Thread Tools

OCR program needed

 
 
richard
Guest
Posts: n/a
 
      12-07-2010
http://1littleworld.net/images/test001.jpg

Tried a program called "Simple OCR" and got a bunch of garbage as a result.
I need a decent OCR that will convert the text on this image to actual
usable text. I know that the special items such as encirculated numbers and
those in a "star" will be difficult to convert, but that's not a big deal.
I'm mainly after the actual text.

Freeware ok but I can spend a few bucks for a better program.
 
Reply With Quote
 
 
 
 
Mike Easter
Guest
Posts: n/a
 
      12-07-2010
richard wrote:
> http://1littleworld.net/images/test001.jpg


That is a very poor reproduction of something.

> Tried a program called "Simple OCR" and got a bunch of garbage as a result.
> I need a decent OCR that will convert the text on this image to actual
> usable text.


That is not going to happen. You need a much better image somehow. That
looks like someone made a scan without sufficient resolution or image
enhancement and then saved it as a lossy format such as .jpg.

I would scan it at about 600 dpi in grayscale and then tweak its
contrast and gamma with something like IrfanView so that everything
looks as if it were sharp black and white. If you have access to the
original, you should be able to make it look better than the original,
sorta. At least the background will be white instead of beige and the
print will be black instead of brownish and the characters will be
something besides blurry.

You might be able to do it at 300 dpi.

> I know that the special items such as encirculated numbers and
> those in a "star" will be difficult to convert, but that's not a big deal.
> I'm mainly after the actual text.
>
> Freeware ok but I can spend a few bucks for a better program.


Your software is the least of your problems. The problem is the image
quality.

You can't make a silk purse out of a sow's ear -or- garbage in, garbage out.


--
Mike Easter
 
Reply With Quote
 
 
 
 
richard
Guest
Posts: n/a
 
      12-07-2010
On Tue, 07 Dec 2010 10:48:31 -0800, Mike Easter wrote:

> richard wrote:
>> http://1littleworld.net/images/test001.jpg

>
> That is a very poor reproduction of something.
>
>> Tried a program called "Simple OCR" and got a bunch of garbage as a result.
>> I need a decent OCR that will convert the text on this image to actual
>> usable text.

>
> That is not going to happen. You need a much better image somehow. That
> looks like someone made a scan without sufficient resolution or image
> enhancement and then saved it as a lossy format such as .jpg.
>
> I would scan it at about 600 dpi in grayscale and then tweak its
> contrast and gamma with something like IrfanView so that everything
> looks as if it were sharp black and white. If you have access to the
> original, you should be able to make it look better than the original,
> sorta. At least the background will be white instead of beige and the
> print will be black instead of brownish and the characters will be
> something besides blurry.
>
> You might be able to do it at 300 dpi.
>
>> I know that the special items such as encirculated numbers and
>> those in a "star" will be difficult to convert, but that's not a big deal.
>> I'm mainly after the actual text.
>>
>> Freeware ok but I can spend a few bucks for a better program.

>
> Your software is the least of your problems. The problem is the image
> quality.
>
> You can't make a silk purse out of a sow's ear -or- garbage in, garbage out.


I did not create the image to begin with. So I do not have the original
scan to work with.
My guess is these images were taken from old newspaper clippings.
 
Reply With Quote
 
Jordon
Guest
Posts: n/a
 
      12-07-2010
richard wrote:
> http://1littleworld.net/images/test001.jpg
>
> Tried a program called "Simple OCR" and got a bunch of garbage as a result.
> I need a decent OCR that will convert the text on this image to actual
> usable text. I know that the special items such as encirculated numbers and
> those in a "star" will be difficult to convert, but that's not a big deal.
> I'm mainly after the actual text.
>
> Freeware ok but I can spend a few bucks for a better program.


If you can't read half of it, how's OCR suppose to?
 
Reply With Quote
 
Mike Easter
Guest
Posts: n/a
 
      12-07-2010
richard wrote:
> Mike Easter wrote:


>> That is a very poor reproduction of something.


> I did not create the image to begin with. So I do not have the original
> scan to work with.
> My guess is these images were taken from old newspaper clippings.


Since you can read that with your eyeballs, then the fastest way to get
the job done is eyeball input to keyboard output, using the human
eyeball and brain to perform 'optical character recognition' and output
that to a digital result with whatever application is the best.

Probably do it with a table in a word processor or spreadsheet. It looks
like there are about 9 or 10 fields per entry.

Personally, I would try to improve the image a little, but you aren't
going to be able to do very much. You can make it into black and white
from grayscaling it and tweaking the contrast and gamma and you will be
able to make it a little bit sharper, but not very much. You might
enlarge the pic somewhat, but not too much because the lossy format will
cause it to start 'coming to pieces' or pixelating,

That will make it easier/faster to read with your eyeballs. There is no
way that an OCR program is going to be able to speed that up for you,
because it is going to have way too many errors. Converting it manually
is more efficient because you will be going directly from recognizing
the character/s to keyboarding them as opposed to trying to fix a great
big mess from some OCR.

On this kind of job, it is much much better to be a touch typist,
because you can focus on the target rather than having to move your eyes
to the keyboard a lot. The eyeball range should be from graphic to
spreadsheet table and back to the graphic.


--
Mike Easter
 
Reply With Quote
 
Mike Easter
Guest
Posts: n/a
 
      12-07-2010
Mike Easter wrote:

> Probably do it with a table in a word processor or spreadsheet. It looks
> like there are about 9 or 10 fields per entry.


There is way more information for each entry than you need. (ISTM
itseemstome) It doesn't matter what the song's ranking was that week or
the week before or the week before that.

Actually all you need is the name of the song and the artist and maybe
or maybe not the label.


--
Mike Easter
 
Reply With Quote
 
Mike Easter
Guest
Posts: n/a
 
      12-07-2010
Mike Easter wrote:

> Actually all you need is the name of the song and the artist and maybe
> or maybe not the label.


El Paso Marty Robbins
Why Frankie Avalon
The Big Hurt Toni Fisher
Running Bear Johnny Preston
Way Down Yonder in New Orleans Freddie Carson
Heartaches by the Number Guy Mitchell
It's Time to Cry Paul Anka
Among My Souvenirs Connie Francis
Pretty Blue Eyes Steve Lawrence
Go, Jimmy, Go Jimmy Clanton
We Got Love Bobby Rydell
Mack the Knife Bobby Darin
The Village of St. Bernadette Andy Williams
Uh! Oh! (Part II) The Nutty Squirrels
Sandy Larry Hall
Hound Dog Man Fabian
Scarlet Ribbons The Browns
Smokie (Part II) Bill Black's Combo
Not One Minute More Della Reese
Friendly World Fabian
In The Mood Ernie Fields
The Little Drummer Boy Harry Simeone Chorale
You Got What It Takes Marv Johnson
First Name Initial Annette Funicello
So Many Ways Brook Benton
Oh, Carol Neil Sedaka
Be My Guest Fats Domino
Teardrop Santo and Johnny
Teen Angel Mark Dinning
Come Into My Heart Lloyd Price
Mr. Blue Fleetwoods
I Wanna Be Loved Ricky Nelson
Danny Boy Conway Twitty



--
Mike Easter
 
Reply With Quote
 
Jordon
Guest
Posts: n/a
 
      12-07-2010
Mike Easter wrote:
> Mike Easter wrote:
>
>> Actually all you need is the name of the song and the artist and maybe
>> or maybe not the label.

>
> El Paso Marty Robbins


[snipped]

I can not believe you did that.


 
Reply With Quote
 
Mike Easter
Guest
Posts: n/a
 
      12-07-2010
Jordon wrote:
> Mike Easter wrote:


>>> Actually all you need is the name of the song and the artist and maybe
>>> or maybe not the label.

>>
>> El Paso Marty Robbins

>
> [snipped]
>
> I can not believe you did that.


That's the old 'one thing leads to another' syndrome.

At first I was just curious about how to tweak the graphics to make
things more legible. Then when I realized that the quality of the
original .jpg was just too poor to tweak up to clarity, it became a
combination of my familiarity with the artists of 1959 hits and trying
to read the blurry artist's names that I couldn't remember.

Gradually it turned into one of those 'moods' like finishing some
crossword puzzle that you've nearly completed. By the time I could read
all of them, it only took a few minutes to type the result, which was
the easiest part.


--
Mike Easter
 
Reply With Quote
 
Jordon
Guest
Posts: n/a
 
      12-07-2010
Mike Easter wrote:
> Jordon wrote:
>> Mike Easter wrote:

>
>>>> Actually all you need is the name of the song and the artist and maybe
>>>> or maybe not the label.
>>>
>>> El Paso Marty Robbins

>>
>> [snipped]
>>
>> I can not believe you did that.

>
> That's the old 'one thing leads to another' syndrome.
>
> At first I was just curious about how to tweak the graphics to make
> things more legible. Then when I realized that the quality of the
> original .jpg was just too poor to tweak up to clarity, it became a
> combination of my familiarity with the artists of 1959 hits and trying
> to read the blurry artist's names that I couldn't remember.
>
> Gradually it turned into one of those 'moods' like finishing some
> crossword puzzle that you've nearly completed. By the time I could read
> all of them, it only took a few minutes to type the result, which was
> the easiest part.


I'll have to remember that when I get around to creating
my thesis on quantum mechanics. Maybe a plea for help here
would get you to write it for me?
 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
OCR Program? Sylvia M Computer Support 10 04-21-2006 06:15 AM
OCR yoshitha ASP .Net 0 11-19-2004 04:08 AM
How to Implement OCR Technology in Asp.net =?Utf-8?B?VmVua2F0Tg==?= ASP .Net 0 02-06-2004 05:06 AM
Urgent! OCR SDK and API or OCR COM integration help Anjali Lourda ASP .Net 0 02-04-2004 06:52 PM
Re: OCR Support Bill Devine ASP .Net 1 08-23-2003 01:29 PM



Advertisments