This Python script extracts email addresses from an image. It receives the path to the input image and the path to the output text file as parameters.
python3 -m venv venv
- create venv. venv/bin/activate
- activate venvpip install -r requirements.txt
- install required libs
NOTE: you can use emages from extractor package. Path: extractor/test-images
python3 main.py --input <path_to_input_image> --output <path_to_output_text_file> example: python3 main.py --input extractor/test-images/image-2.png --output output.txt
- In this project I didn't use
pytesseract
library, because I couldn't reach the precision as witheasyocr
. easyocr
lib is multilingual. Project repoeasyocr
lib Supoorted languages- The image preprocesses before extracting via
opencv
lib - it thresholds (binarizes). This should be improved - finetune easyocr's readtext
- finetune easyocr's model
- extract text/emails with LLM