vik-ma / screenshot-ocr

Desktop application that lets the user extract text from images by just marking a section of the screen, instead of having to load an image file. Serves as a front-end for the Tesseract OCR Engine.

Screenshot OCR

Screenshot OCR is a desktop GUI frontend for Tesseract OCR Engine made for Windows. The application allows the user to select a snippet of their screen, automatically execute Tesseract and extract the text from the snippet.

Features

• Select a snippet of your screen to perform OCR on, without needing to save or load an image
• Automatically copy OCR output to clipboard
• Perform OCR on a local file
• Keyboard shortcuts (Can be disabled)
• Select language to perform OCR on from a list of installed languages in Tesseract
• Add additional languages as parameters to OCR
• Save default language between uses
• Save combinations of base language + additional language parameters between uses
• Edit output text in application
• Option to save OCR output as .txt file (Also option to select destination folder)
• Option to save screen snippet as .png file (Also option to select destination folder)
• Save user configuration between uses

Requirements

Screenshot OCR requires Python to run (Version 3.10 or newer is recommended) and the following packages:

• PyQt5
• pytesseract

Everything can be installed from either requirements.txt or the Pipfile.

For the program to work you also need Tesseract OCR Engine for Windows installed on your system.

How To Use

The application will look for a Tesseract executable in 'C:\Program Files\Tesseract-OCR\tesseract.exe'. If your installation of Tesseract is located elsewhere, you can manually select it when first running the program.

Language And Additional Language Parameters

Before performing OCR, select the language you want to read from the first list-box from the left. If you want to add additional languages, you can select them from the third list-box and then click the 'Add Language' button below the list-box. All additional languages now appear in the second list-box. To remove an additional language, select the language and click the 'Remove Language' button below the second list-box.

Perform OCR On Screen Snippet

Click on the 'Take Snippet' button (or just press 'S' on your keyboard) to bring up the snippet functionality. Hold down the left mouse-button and draw a rectangle around the text your want to read, and when you let go, the application will automatically perform OCR on the selected area. The output will then be pasted into the text-field in the application.

You can press 'Esc' to exit the snippet functionality without performing OCR on anything.

The snippet functionality works on multiple monitors, but may not completely cover everything if the monitor's resolutions and positions differ wildly from each other. Move the text you want to OCR onto your main monitor if this is an issue.

Perform OCR On Local File

Click on the 'Read File' button (or press 'F' on your keyboard) to select an image file on your computer to perform OCR on. The output will be pasted into the text-field in the application once the file has been selected.

Saved Configuration

User configuration is stored in a 'config.ini' file that gets automatically created in the same directory as 'screenshot_ocr.py' when it's run for the first time.

To change the default language to perform OCR on, select the language by clicking on it in the first list-box from the left, and then click the 'Set Default' button below the list-box. That language will now be automatically selected every time the application is run.

Saved Language Combinations

You can save a combination of additional language parameters by clicking the 'Save Language Combo' button. This will save the currently selected language along with all additional languages present in the second list-box from the left. Multiple language combinations can be saved and retrieved from the dropdown menu underneath 'Saved Language Combos'.

By clicking the 'Set Combo As Default' button, you can set the currently selected saved language combination to be automatically selected every time the application is run.