This is a demo of approaches that extract objects from images. The first version of this app is able to extract textual data from driving license and save it as an excel file / demonstrate results in a simple user interface
Using Docker:
docker build . -t "cv:v1"
docker run -p 4200:4200 "cv:v1"
By default you can access the app here: http://localhost:4200/
-
Pytesseract can be replaced by a custom model
-
Rule-based approach of parsing concrete entities can be improved through custom NER model creation