/ocr-gui

Primary LanguagePython

An OCR GUI powered by PaddleOCR.

It reads the image from clipboard, so another screenshot tool such as Snipaste is required.

Install Dependecies

inside or outside a virtual environment:

pip install -r requirements.txt

Run

python tk.py

Pack to exe(Build)

There are 2 approaches:

  1. pyinstaller command:
pyinstaller --noconfirm --onefile --windowed --hidden-import "imghdr" --hidden-import "imgaug" --hidden-import "pyclipper" --collect-data "paddle" --collect-all "paddleocr"  "C:\Users\naive\ocr-gui\tk.py"

  1. auto-py-to-exe config json:

    1. install auto-py-to-exe: pip install auto-py-to-exe

    2. run auto-py-to-exe

    3. choose the apte.json in Settings > Import Config From Json File

Web Update

Since AI models are large, maybe a better way to hold the application is in the form of web application.

I have made a web-equivalent of tk.py. check it by

python flask_backend.py