Is there a GUI for OCRmyPDF?

[ownedbycats]

https://wiki.mobileread.com/wiki/OCRmyPDF

I am... fairly bad with using the command-line. Does anybody know if there was any sort of GUI made for this?

[willus]

For any particular platform / operating system? Looks like OCRmyPDF is largely targeted at linux but can install on OS/X and Windows with some third party support (e.g. Python, Homebrew/Cygwin...). There are a lot of other OCR apps if you find OCRmyPDF difficult to use. If you have an iPhone, for example, there is Elucidate which also uses the Tesseract OCR engine.

[ownedbycats]

I somehow managed to completely skip over the fact that there's no real Windows implementation

[Quoth]

It's really for tesseract anyway. Even on Linux you might use something else with Tesseract!

You can run Linux for free, either on a VM (Openbox is free on Windows 10 and recommended MS solution for XP and Win7 on Win10), or USB stick or dual boot, or ditch windows (me entirely in Jan 2017, but I have a clone of my 2002 XP laptop on OpenBox VM on Linux and Office 2003 on WINE).

[JSWolf]

There's no good program for OCRing PDF. It's not possible.

[j.p.s]

Originally Posted by JSWolf

There's no good program for OCRing PDF.

That might well be true.

Originally Posted by JSWolf

It's not possible.

Prove it or retract.

[JSWolf]

Originally Posted by j.p.s

That might well be true.



Prove it or retract.

If it was possible with convert PDF to something else easily that doesn't need a lot of fixing, it would have already been done.

[j.p.s]

Originally Posted by JSWolf

If it was possible with convert PDF to something else easily that doesn't need a lot of fixing, it would have already been done.

And that is why we already have cheap fusion energy for our cities on Mars.

You have proved nothing. Prove your assertion, retract, or rephrase to be true.

(You have been told that makking an assertion does not establish a fact and it certainly does not constitute proof.)

[Quoth]

Originally Posted by JSWolf

There's no good program for OCRing PDF. It's not possible.

it depends on the PDF image quality. On a good day it can beat a copy-typist.

That's too general a statement.

You certainly need to proof & edit the output.

[salamanderjuice]

Originally Posted by JSWolf

If it was possible with convert PDF to something else easily that doesn't need a lot of fixing, it would have already been done.

It really depends what you're using the OCR'd text to do. To make an ePub? Yeah it'll need some work. To search the text of a bunch of PDFs? It can do a pretty good job.

[ownedbycats]

I am now very confused. I just wanted to add a text layer to some scanned booklets I had for search purposes.

[salamanderjuice]

Tesseract is an open source OCR engine that a lot of programs use, OCRmyPDF included. There's a number of programs to do OCR for free using it like gImageReader that works on Windows with a GUI. I don't think gImageReader will embed the text in a layer like OCRmyPDF but it will let you get a text document out at least.

JSWolf seems to think OCR on PDF is useless but he's wrong.

If you have Windows 10/11 it's also not that hard to get OCRmyPDF working under the Windows Subsystem for Linux (WSL).

There's also some websites that will do it like https://www.sandwichpdf.com/

[willus]

Originally Posted by ownedbycats

I am now very confused. I just wanted to add a text layer to some scanned booklets I had for search purposes.

If you cannot get ocrmypdf working in Windows or are not comfortable with the command line, you can try k2pdfopt. It will unfortunately re-render each page as a new bitmap (this can be good if you want to change the resolution of the original PDF, or it can be bad if the original PDF uses better a better compression technique than k2pdfopt). If using the MS Windows GUI, select "copy" for the conversion mode and the check the "OCR (Tesseract)" box.

If you dig deeper there are options to adjust contrast, gamma correction, and output resolution.