a web service which can take an uploaded image and find any letters in it
Run the following command to launch the web service:
docker-compose up
then open http://127.0.0.1:8000/recognization/
Run the following command:
python manage.py test
The ocr test will evaluate the edit distance between the actual string and the string we get from image. If the edit distance is less than one-tenth of the actual string, we consider it an successful ocr.
-
pytesseract~=0.3.7: python wrap for tesseract, an OCR package written in C++
-
Pillow~=8.1.0
-
Django~=3.1.5
-
editdistance~=0.5.3: evaluate the edit distance between the actual string and the string we get from the image.
-
opencv-python: preprocess the image, perform a threshold to segment the foreground from the background,
or apply a median blur to reduce noise.