Carriage return after each word

[pascal 06]

Hello everybody,

I'm using Tesseract with GScan2pdf under Linux.

When I do an OCR on a document, Tesseract recognition is quite good however it put a carriage return after each word. It is very annoying !

Here an example on what it does:

1. Scanned document in gscand2pdf:
   
2. Text recognition in gscan2pdf:
   
3. Generated pdf opened with Okular; copy selected text:
   
4. Pasted text:
   Conformément
    à l’article
    12 du
    Règlement du Fonds, le Fonds a procédé à sa deuxième
    distribution.
   Ce deuxième
    remboursement
    de capital
    s'élève
    à
    un
    montant
    de
    5.74
    €
    par part,
    soit
    5,74 % du nominal
   investi.
   

Here information about my system:

pascal@pascal-Latitude-5580:~$ lsb_release -a 
No LSB modules are available.
Distributor ID: Ubuntu

Description: Ubuntu 24.04.3 LTS 

Release: 24.04 

Codename: noble
pascal@pascal-Latitude-5580:~$ tesseract --version 

tesseract 5.3.4

 leptonica-1.82.0

 libgif 5.2.1 : libjpeg 8d (libjpeg-turbo 2.1.5) : libpng 1.6.43 : libtiff 4.5.1 : zlib 1.3 : libwebp 1.3.2 : libopenjp2 2.5.0 

Found AVX2 

Found AVX 

Found FMA Found SSE4.1 

Found OpenMP 201511 

Found libarchive 3.7.2 zlib/1.3 liblzma/5.4.5 bz2lib/1.0.8 liblz4/1.9.4 libzstd/1.5.5 

Found libcurl/8.5.0 OpenSSL/3.0.13 zlib/1.3 brotli/1.1.0 zstd/1.5.5 libidn2/2.3.7 libpsl/0.21.2 (+libidn2/2.3.7) libssh/0.10.6/openssl/zlib nghttp2/1.59.0 librtmp/2.3 OpenLDAP/2.6.7

Any idea ?

Thanks a lot in advance for your help.

[pascal 06]

Hello,

so somebody suggested me to do an OCR directly from command line. So using the example above, I did:

~~~

tesseract test.tif test.txt -l fra

~~~

and then the result:

>Conformément à l’article 12 du Règlement du Fonds, le Fonds a procédé à sa deuxième distribution.
>Ce deuxième remboursement de capital s'élève à un montant de 5.74 € par part, soit 5,74 % du nominal
>
>investi.

Do you have any idea what's going on?
Thank you in advance for any help you can provide,
Pascal

[Tom Morris]

Salut Pascal,

I'm glad that you were able to determine that Tesseract is working correctly.

On Wednesday, September 17, 2025 at 6:57:45 AM UTC-4 pascal....@gmail.com wrote:

Do you have any idea what's going on?

Since you are working with two different applications, which folks in this forum are

unlikely to have much knowledge of: 1) gscan2pdf, which uses/embeds Tesseract,

and 2) Okular, which I'm guessing is a PDF viewer.

There are a number of areas where things could go awry, including the way the PDF

is constructed and the way the text is selected and formatted on the clipboard.

I suspect that the text is being split into multiple text blocks and that each of those

text blocks is getting a new line added for "free" at the end. Where in the processing

chain this is happening isn't clear.

If your goal is simply to get the best rendition of the text, it sounds like you've 

discovered what is needed. If you want to get that specific combination of 

programs to work better, you're probably going to need to address it with

whoever supports them.

Bonne chance!

Tom

 

[pascal 06]

Bonsoir Tom,

je suppose que tu es francophone :)

Merci pour ta réponse !

Je vais continuer en anglais pour une meilleure compréhension pour les autres personnes :)

So,

Thanks a lot for your reply.

Indeed, Okular is a pdf reader.

I agree most of the things. However, I think that things already goes wrong in gscan2pdf OCR. As you wrote "I suspect that the text is being split into multiple text blocks and that each of those text blocks is getting a new line added for "free" at the end."

I opened a support request here:
https://sourceforge.net/p/gscan2pdf/support-requests/70/

Hope I'll get help :)

In conclusion, for the moment the most important thing is I can do text search in many scanned document thanks to Tesseract.

Encore merci pour ton aide,

Pascal