Image to text (OCR feature)

[psychoSherlock]

OCR ( Optical Character Recognition ) in pywhatkit

PyWhatKit is great because of its ability to simple every python tasks. I would like to add a feature that converts hand written texts to machine readable texts (OCR)

Basically a user can provide any image and pywhatkit would extract all the texts out of the image. I thought this might be a great feature to have.

We could use any python modules that supports OCR, but the one I like most is pytesseract as its very accurate and uses google's OCR engine.

[aaryanrr]

Hey @psychoSherlock , It would be a good feature, we'll see what can be done.

[aaryanrr]

Hey @psychoSherlock, OCR can be an useful feature but most of the offline solutions that we found are slow and it'll require some heavy dependencies. We really want this module to be lightweight. So we are not planning on including this feature. Thanks for your valuable suggestion!

[codewithpom]

@aaryanrr Just use an API that we will make in flask or fast-api

[aaryanrr]

That would require a better server.

[codewithpom]

why ?

[psychoSherlock]

why ?

I think he's right (@aaryanrr)making a server just for OCR detection may arise many security issues, which is not something we could handle, and also it costs much to buy a server that's faster, resourcefull for OCR detection and with a high network bandwidth

[codewithpom]

@psychoSherlock Just host on repl.it or heroku and use a module with Google OCR and please tell me what are the security issues. :slightly_smiling_face:

[psychoSherlock]

@psychoSherlock Just host on repl.it or heroku and use a module with Google OCR and please tell me what are the security issues. :slightly_smiling_face:

Pywhatkit is currently just a non profit small community right?, so I belive it would be hard for them to host and manage these servers. As @aaryanrr said it's better if we kept it light weight

Why bother for hosting if paid APIs are already available, read here.

Or if u wanna use a offline version, use pytesseract 🙃

[codewithpom]

Python-tesseract uses Google OCR hence even it uses internet.

[aaryanrr]

@psychoSherlock Just host on repl.it or heroku and use a module with Google OCR and please tell me what are the security issues. 🙂

Pywhatkit is currently just a non profit small community right?, so I belive it would be hard for them to host and manage these servers. As @aaryanrr said it's better if we kept it light weight

Why bother for hosting if paid APIs are already available, read here.

Or if u wanna use a offline version, use pytesseract 🙃

Yep correct. And offline use requires the installation of the Tesseract OCR Engine and adding it to the path so that is some extra steps required.