image processing

17 views
Skip to first unread message

Aniket Gadge

unread,
Apr 16, 2021, 7:47:36 AM4/16/21
to django...@googlegroups.com
hello everyone,
              my college project is Ai based in that i want to store the student details to a database, The form is a scan document in image format and i want to retrieve that data like name , age, phone number , etc. how can i retrieve data which is one user form?


                                                                      

Kasper Laudrup

unread,
Apr 16, 2021, 7:56:02 AM4/16/21
to django...@googlegroups.com
What you're looking for is called OCR (Optical Character Recognition).
There seems to be library for Python called Tesseract which should be
simple to integrate with Django. Here's a tutorial:

https://towardsdatascience.com/create-simple-optical-character-recognition-ocr-with-python-6d90adb82bb8

I don't have any experience with this library or OCR in general, but
hopefully this should guide you in the right direction.

Good luck with your project.

Kind regards,

Kasper Laudrup

OpenPGP_signature

Aniket Gadge

unread,
Apr 16, 2021, 8:04:23 AM4/16/21
to django...@googlegroups.com
Thanks ,
     but i want form data which is in the image format and ocr library is not capable of giving me any perticular data it gives me all the test which is on the image and as well it also not gives me correct image text many time


--
You received this message because you are subscribed to the Google Groups "Django users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to django-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/django-users/0b8549df-4f8b-94e6-ec79-8ba850b9ca98%40stacktrace.dk.

Kasper Laudrup

unread,
Apr 16, 2021, 8:11:51 AM4/16/21
to django...@googlegroups.com
On 16/04/2021 14.00, Aniket Gadge wrote:
> Thanks ,
>      but i want form data which is in the image format and ocr library
> is not capable of giving me any perticular data it gives me all the test
> which is on the image and as well it also not gives me correct image
> text many time
>

I assume you mean text and not test?

You have to write some code to extract the relevant parts of the text.
No library is going to do that for your specific use case.

There also is no way you can be sure to get the correct text all the
time, but you can probably tune the algorithm used for text recognition
in different ways.

Kind regards,

Kasper Laudrup

OpenPGP_signature

Aniket Gadge

unread,
Apr 19, 2021, 2:36:39 AM4/19/21
to django...@googlegroups.com
Still not get it
Reply all
Reply to author
Forward
0 new messages