REDDIT 原始帖子
Looking for an OCR tool
I need an OCR tool that should convert images to word and works completely offline and doesn't leak data and keeps documents private. If it is free that's great but paid will work too. (Also I have windows 10 if that helps.)
I need an OCR tool that should convert images to word and works completely offline and doesn't leak data and keeps documents private. If it is free that's great but paid will work too. (Also I have windows 10 if that helps.)
已收录讨论
PyTesseract Anything else is a pig in lipstick. Don't pay for things that are free, or use gimmicky tools that are based on the original free variant. Most of the things recommended in this thread are powered by PyTesseract under the hood anyway. https://pypi.org/project/pytesseract/
Thanks for your help chatgpt
Thanks I will try that
The original is technically https://github.com/tesseract-ocr/tesseract That’s just a python wrapper
generally speaking OCR on CPU is very slow, if you have a decent GPU the OCR speed and quality will much better. I can show you how to write some simple python code for OCR.
A python wrapper that makes it much easier to install on various systems.
Okay thanks for advice
I really appreciate your help bro But my work has been done Still thanks alot
There's a great chance that you have one of the best tools installed in your win10 pc Just paste your image in any onenote document, then right click the image and select 'copy text' ... and that's it
NAPS2 free scanner software
what kind GPU do you have?
If you have quite a few images, automate it using Powershell. There are lots of tutorials and pre-written scripts and modules that already do this using the Tesseract OCR engine. Just google "Powershell OCR" and the first page contains links to the entire solution. Its surprisingly easy to do, and you can let it run overnight if you have a ton of images, and wake up to hundreds of text files.
ocrmypdf... but you gotta know basic command line stuff.
PDFGear
Check naps2
look for an offlines ocr app theyre more private since everything stays on your computer
I cleaned up a bunch of images (batch processed b/w, crop, sharpened levels etc. with Ifranview), of a typed out book a client gave me and processed the ocr (hilarious, in my view) with Microsoft Photos' OCR tool. It came out perfectly, ran offline on my old win10 laptop. I cut and pasted the entire document page by page in LibreOffice, set up a few automated find/replace tools to clean up the broken formatting, and then read through the entire thing to double check grammar/spelling and formatting. It was an ~120 page book, took ~7 hours start to finish. I was surprised that Microsoft did something useful, for once. I'm not sure about other peoples' needs and experiences. In this case, it worked out.
Karja has a great OCR app (it's called Scanner). Plus it synergises with the Notepad app (you can export your OCR results there) from which you can use your extracted content with your local AI. Karja is completely private (your scanned medical or financial documents don't end up in someone else's cloud) and works offline forever.
Ollama OCR
KillerPDF now works via CLI for batch jobs and scripting.
Is there a plug and play option of this so I can just download and start it?
KillerPDF has OCR and can be used from command line for scripting and batch jobs
KillerPDF is worth a look
这条评论已被删除。
这条评论已被删除。