Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #185755 > unrolled thread

OCR

Back to article view | Back to linux.debian.user

Contents

  OCR "Stephen Grant Brown" <steve.brown_nbn@iinet.net.au> - 2017-08-23 00:40 +0200
    Re: OCR Doug <dmcgarrett@optonline.net> - 2017-08-23 01:10 +0200
      Re: OCR MENGUAL Jean-Philippe <mengualjeanphi@free.fr> - 2017-08-23 06:40 +0200
    Re: OCR Miles Fidelman <mfidelman@meetinghouse.net> - 2017-08-23 04:10 +0200
    Re: OCR Jerome Flesch <jflesch@gmail.com> - 2017-08-23 08:30 +0200
      Re: OCR Jerome Flesch <jflesch@gmail.com> - 2017-08-23 09:30 +0200
        Re : Re: OCR MENGUAL Jean-Philippe <mengualjeanphi@free.fr> - 2017-08-23 10:00 +0200
      Re: OCR MENGUAL Jean-Philippe <mengualjeanphi@free.fr> - 2017-08-23 10:20 +0200
    Re: OCR Joe <joe@jretrading.com> - 2017-08-23 10:00 +0200
      Re : Re: OCR MENGUAL Jean-Philippe <mengualjeanphi@free.fr> - 2017-08-23 10:00 +0200
      Re: OCR Michael Grant <mgrant@grant.org> - 2017-08-23 10:20 +0200

#185755 — OCR

[Multipart message — attachments visible in raw view] — view raw

Hi All,
What is the best OCR package to use to scan the receipts given immediately after making a purchase?
Yours Sincerely
Stephen Grant Brown 

[toc] | [next] | [standalone]

#185758

[Multipart message — attachments visible in raw view] — view raw

On 08/22/2017 05:22 PM, Stephen Grant Brown wrote:
> Hi All,
> What is the best OCR package to use to scan the receipts given 
> immediately after making a purchase?
> Yours Sincerely
> Stephen Grant Brown
I wish there was a good OCR in Linux, but if there is, I don't know of 
it. You need Windows. Abbyy Fine Reader is excellent. When I need an 
OCR, I boot Windows--just about
the only time I do. (The other time is when I need good audio copy from 
LP records--DAK.)

--doug

[toc] | [prev] | [next] | [standalone]

#185765

Hi,


Many great solutions exist now for OCR on Linux.

1. Free software: gimagereader (uses Tesseract), works all right, as
well as gocr

2. Other free solution: Hypra developed Ocrizer. If you install it, you can:

- from a shortcut, running the scanning process, the characters
recognition, and opening Writer. The shortcut is integrated
automatically in Hypra systems, using MATE and Compiz. As you probably
not installed it, you will need to implement the binding by hand. The
command to add to your desktop environment and to give a binding would
be: sh -c "/usr/bin/ocrizer -s -d ~/Documents"

- if using MATE, you can install 
mate-accessibility-ocr-integration-caja and a command in the context
menu will enable you to OCR a n image file (eg a PDF shipped on the
Internet), via different means.


Note that this tool is based on Tesseract.


3. Paying solution:

a) Abbyy Finereader (proprietary): for 150 euros, Abbyy provides a
commandline tool. Install it, run it via a script you create or
manually, and you can use Finereader engine in Linux. The license is a
key, you can get online, it enables you to be run on One machine and you
can OCr 12000 pages per year (the counter reloads each year).


If you buy it but want to have a script, you can install it, then
install ocrizer-finereader from Hypra. It will bring the features I
described above, but based on the Finereader engine instead of Tesseract.


b) Finally, if you want to support a social-impact company and get
support to implement your OCR installation, and if you want to bypass
the anual 12000 pages limitation, you can buy Finereader to Hypra: just
a mail, then a payment by transfer and they send you the license key.
The price is 180 euros, because the anual 12000 pages is missing.


To get Hypra repos (for suggestions 2 and 3 a):

http://hypra.fr/Developpement-72.html


Note that only Jessie works now, devs for Stretch are in progress. But
Hypra can assist you installing Finereader and Ocrizer on another
release of Debian (it's paying but possibe).


Hope it helps.



Cheers


Jean-Philippe MENGUAL

[toc] | [prev] | [next] | [standalone]

#185762

[Multipart message — attachments visible in raw view] — view raw

Probably CamScanner on a smartphone.


On 8/22/17 3:22 PM, Stephen Grant Brown wrote:
> Hi All,
> What is the best OCR package to use to scan the receipts given 
> immediately after making a purchase?
> Yours Sincerely
> Stephen Grant Brown

-- 
In theory, there is no difference between theory and practice.
In practice, there is.  .... Yogi Berra

[toc] | [prev] | [next] | [standalone]

#185767

Hello,

Since it may be relevant, I'm going to do some advertisement for my
own software : https://openpaper.work .

I hope it can help you,



2017-08-23 0:22 GMT+02:00 Stephen Grant Brown <steve.brown_nbn@iinet.net.au>:
> Hi All,
> What is the best OCR package to use to scan the receipts given immediately
> after making a purchase?
> Yours Sincerely
> Stephen Grant Brown

[toc] | [prev] | [next] | [standalone]

#185768

[Multipart message — attachments visible in raw view] — view raw

It's based on Tesseract. It can also use Cuneiform, but results are not as
good.

2017-08-23 8:55 GMT+02:00 MENGUAL Jean-Philippe <mengualjeanphi@free.fr>:

> Hi,
>
>
> I will test it, happy to know a new tool. Is it based on tesseract as well
> or is it your own engine?
>
>
> Best regards,
>
>
>
> Le 23/08/2017 à 08:21, Jerome Flesch a écrit :
>
> Hello,
>
> Since it may be relevant, I'm going to do some advertisement for my
> own software : https://openpaper.work .
>
> I hope it can help you,
>
>
>
> 2017-08-23 0:22 GMT+02:00 Stephen Grant Brown <steve.brown_nbn@iinet.net.au> <steve.brown_nbn@iinet.net.au>:
>
> Hi All,
> What is the best OCR package to use to scan the receipts given immediately
> after making a purchase?
> Yours Sincerely
> Stephen Grant Brown
>
>
> --
> [image: Logo Hypra] JEAN-PHILIPPE MENGUAL
> DIRECTEUR TECHNIQUE ET QUALITÉ
> 102, rue des poissonniers, 75018, Paris
> Tel : +331 84 73 06 61 <+33184730661> Mob : +336 76 34 93 37
> <+33676349337>
> jpmengual@hypra.fr
> www.hypra.fr
> [image: Facebook Hypra] <https://www.facebook.com/hyprasoftware/> [image:
> Twitter Hypra] <https://twitter.com/Hypra_> [image: Linkedin
> Jean-Philippe]
> <https://fr.linkedin.com/in/jean-philippe-mengual-800133135>
>
>

[toc] | [prev] | [next] | [standalone]

#185770 — Re : Re: OCR

Ok interesting, I will test and do feedbacks. The people I work with is interesting for testing because blind, as myself. Then they need a good quality OCR to read common image files. When I searched two years ago, I had not found very accessible an performant software (gimagereader is in GTK but misses manems; etc). Hence my look at Abbyy Linux sdk. But always interested if I can get rid of it.

I plan, once having human resources, to package Ocrizer in debian repo. What about your utility?

Best regards,


Jean-Philippe MENGUAL

HYPRA, progressons ensemble

Tél.: 01 84 73 06 61

Mail: contact@hypra.fr

Site Web: http://www.hypra.fr

----- Jerome Flesch <jflesch@gmail.com> a écrit :
> It's based on Tesseract. It can also use Cuneiform, but results are not as
> good.
> 
> 2017-08-23 8:55 GMT+02:00 MENGUAL Jean-Philippe <mengualjeanphi@free.fr>:
> 
> > Hi,
> >
> >
> > I will test it, happy to know a new tool. Is it based on tesseract as well
> > or is it your own engine?
> >
> >
> > Best regards,
> >
> >
> >
> > Le 23/08/2017 à 08:21, Jerome Flesch a écrit :
> >
> > Hello,
> >
> > Since it may be relevant, I'm going to do some advertisement for my
> > own software : https://openpaper.work .
> >
> > I hope it can help you,
> >
> >
> >
> > 2017-08-23 0:22 GMT+02:00 Stephen Grant Brown <steve.brown_nbn@iinet.net.au> <steve.brown_nbn@iinet.net.au>:
> >
> > Hi All,
> > What is the best OCR package to use to scan the receipts given immediately
> > after making a purchase?
> > Yours Sincerely
> > Stephen Grant Brown
> >
> >
> > --
> > [image: Logo Hypra] JEAN-PHILIPPE MENGUAL
> > DIRECTEUR TECHNIQUE ET QUALITÉ
> > 102, rue des poissonniers, 75018, Paris
> > Tel : +331 84 73 06 61 <+33184730661> Mob : +336 76 34 93 37
> > <+33676349337>
> > jpmengual@hypra.fr
> > www.hypra.fr
> > [image: Facebook Hypra] <https://www.facebook.com/hyprasoftware/> [image:
> > Twitter Hypra] <https://twitter.com/Hypra_> [image: Linkedin
> > Jean-Philippe]
> > <https://fr.linkedin.com/in/jean-philippe-mengual-800133135>
> >
> >

[toc] | [prev] | [next] | [standalone]

#185773

[Multipart message — attachments visible in raw view] — view raw

Hi,


I will test it, happy to know a new tool. Is it based on tesseract as
well or is it your own engine?


Best regards,



Le 23/08/2017 à 08:21, Jerome Flesch a écrit :
> Hello,
>
> Since it may be relevant, I'm going to do some advertisement for my
> own software : https://openpaper.work .
>
> I hope it can help you,
>
>
>
> 2017-08-23 0:22 GMT+02:00 Stephen Grant Brown <steve.brown_nbn@iinet.net.au>:
>> Hi All,
>> What is the best OCR package to use to scan the receipts given immediately
>> after making a purchase?
>> Yours Sincerely
>> Stephen Grant Brown
>

-- 
signature_jp_2
Logo Hypra 	JEAN-PHILIPPE MENGUAL
DIRECTEUR TECHNIQUE ET QUALITÉ
102, rue des poissonniers, 75018, Paris
Tel : +331 84 73 06 61 <tel:+33184730661> Mob : +336 76 34 93 37
<tel:+33676349337>
jpmengual@hypra.fr <mailto:jpmengual@hypra.fr>
www.hypra.fr <http://www.hypra.fr/>
Facebook Hypra <https://www.facebook.com/hyprasoftware/> Twitter Hypra
<https://twitter.com/Hypra_> Linkedin Jean-Philippe
<https://fr.linkedin.com/in/jean-philippe-mengual-800133135>

[toc] | [prev] | [next] | [standalone]

#185769

On Wed, 23 Aug 2017 08:22:37 +1000
"Stephen Grant Brown" <steve.brown_nbn@iinet.net.au> wrote:

> Hi All,
> What is the best OCR package to use to scan the receipts given
> immediately after making a purchase? Yours Sincerely

If you mean general retail receipts, are you sure this is a practical
proposition?

I have to look closely at various receipts I get in order to decipher
dates and prices, often with the printed characters broken, and
sometimes made with a dodgy printer.

I keep a web-accessible database of 'important' receipts, but I just
scan them to greyscale PNG and manually type in the goods, date,
supplier and category. Even then, I sometimes have to tweak the scan
parameters to get a (human-)legible result.

I have seen a purpose-made business card scanner, which produced what I
thought were amazing results about ten years ago, particularly given the
infinite layout possibilities of cards. But it was seeing brand-new
cards, with high quality commercial printing, not some wrinkled scrap
of paper from an old thermal dot matrix printer, with the
'time-for-a-new-roll' red stripe down the middle.

Best of luck.

-- 
Joe

[toc] | [prev] | [next] | [standalone]

#185771 — Re : Re: OCR

Oh indeed, this case is very complex. Another possible thing is to train tesseract for this kind fo situations, but it is a heavy process indeed.

regards


Jean-Philippe MENGUAL

HYPRA, progressons ensemble

Tél.: 01 84 73 06 61

Mail: contact@hypra.fr

Site Web: http://www.hypra.fr

----- Joe <joe@jretrading.com> a écrit :
> On Wed, 23 Aug 2017 08:22:37 +1000
> "Stephen Grant Brown" <steve.brown_nbn@iinet.net.au> wrote:
> 
> > Hi All,
> > What is the best OCR package to use to scan the receipts given
> > immediately after making a purchase? Yours Sincerely
> 
> If you mean general retail receipts, are you sure this is a practical
> proposition?
> 
> I have to look closely at various receipts I get in order to decipher
> dates and prices, often with the printed characters broken, and
> sometimes made with a dodgy printer.
> 
> I keep a web-accessible database of 'important' receipts, but I just
> scan them to greyscale PNG and manually type in the goods, date,
> supplier and category. Even then, I sometimes have to tweak the scan
> parameters to get a (human-)legible result.
> 
> I have seen a purpose-made business card scanner, which produced what I
> thought were amazing results about ten years ago, particularly given the
> infinite layout possibilities of cards. But it was seeing brand-new
> cards, with high quality commercial printing, not some wrinkled scrap
> of paper from an old thermal dot matrix printer, with the
> 'time-for-a-new-roll' red stripe down the middle.
> 
> Best of luck.
> 
> -- 
> Joe
> 

[toc] | [prev] | [next] | [standalone]

#185772

I have used expensify.com at work.  They have a free tier.  Scanning
and an app on the phone that you just take a photo of receipts as you
go works well.  If you can limit your expenses to 10 scans a month
it's free.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web