Closed Zio3D closed 1 month ago
This is a bug with the underlying OCR models being used. I am eager to use updated models when they become available to 3rd party developers, but currently the best way to improve would be to use the Text Grab settings to install Tesseract and use those models.
OK, thanks, I'll follow your advice 😃
Describe the bug Does not correctly understand the text of the Ubuntu server console.![Image for the OCR](https://github.com/TheJoeFin/Text-Grab/assets/42025157/a7c6bb50-4024-4342-bc5c-88b6d8725eca)
Where is the bug
To Reproduce Steps to reproduce the behavior: Grab Frame on the image attached
Expected behavior That all characters are recognized correctly as there is no video interference that distorts the text or creates ambiguity about the characters.
Screenshots
Where did you get Text Grab?
Desktop (please complete the following information):
Additional context
If I OCR the attached image (I opened it with Microsoft Paint of Windows 10 Pro) with zoom at 100%, I get the following text:
If I OCR the attached image (I opened it with Microsoft Paint of Windows 10 Pro) with zoom at 200%, I get the following text:
The correct text is:
With the comand:
Get-WindowsCapability -Online | Where-Object { $_.Name -Like 'Language.OCR*' }
, I have these languages installed:Ubuntu console font parameters: FONTFACE="Fixed" FONTSIZE="8x16"