[Juho's AutoHotkey Tutorial #10 Run & Command Prompt] Part 6 - Portable Tesseract & Leptonica (OCR)

preview_player
Показать описание
In this video I walk you thru how to download and use the portable version of Tesseract for Optical Character Recognition (OCR) and Leptonica for image pre-processing to improve the OCR result.

Traineddata for other languages:

Other relevant videos:

Other useful scripts:

To take a look at my AutoHotkey tutorial plan:
• select the dropdown list and click the item See Tutorial Plan Here

AutoHotkey Playlists

Background Music at the end: Ehrling - Lounge

#AutoHotkey #AHK #Automation #Programming #Coding #Course #Tutorial #AHK #Productivity #Efficiency #Office #Shortcut #Hotkey #Macro
Рекомендации по теме
Комментарии
Автор

Juho, congratulations for your videos!
You've made tremendous useful work for AHK and you are helping me a lot!
Did you make a video about GDIP library in full detail?!
I looked at your videos and I see you mentioned you would make it in more than one, but I couldn't find it!
Congratulations again!

saviomilbratz
Автор

It would be good if it can give you X, Y position of a text you are seeking

trilogen
Автор

What if I just want to read some numbers on the screen? Can I not just do it directly without having to save an image?

lolsucks
Автор

Juho let me ask you one thing. I have a script that make screencapture using GDIP, specifically of numeric parameters in a image file, then perform OCR using Tesseract, then turn these numbers into variables, and then use these variables as input for other purposes. The numeric parameters ocupy a relative small part of the screen (few pixels captured). I noticed that once in a while, the OCR reads wrongly the numeric parameters. I've tried to fix this changing the language of Tesseract OCR, and sometimes the wrongly read parameter are then read correctly, but others that were read correctly in the default english language are now wrongly read. Are you aware of a way to improve this reading, either by configuring Tesseract or by any other means?! Training Tesseract seems like complicated to me.
I'm trying to figure that out but I'm not finding any solutions.
Do yoi have something in mind?!
Although the image is clearly a number for a human eye, even after image processing and resizing by leptonica as in the script I've learned from you, it doesn't work.

saviomilbratz
Автор

I couldn't find tesseract script on your website!
Have you updated it?

justforyou
Автор

Hi Juho, do have a code for specific area on the screen for teseract? just like the UWP OCR.

bentukmol
Автор

Hello,
first thank you for your videos, can you tell me what error I have in this code (it does not close the page)
f1::

MyVar := 30 * 60000 ; 30 means minutes and times 60000 gives the time in milliseconds.
Sleep MyVar ; Sleep for 30 minutes.

WinClose, YouTube

return
I await your response as soon as possible.
Thank you

kidsgeschichten