About

This is a hands-on demo which exemplifies a robust method of extracting text from Romanian ID Cards (also known as "buletine") using the Template-Parser service. Unfortunately, the original images used to validate this cannot be published without heavy blurring to prevent disclosure of personal information.

Please find in this article more information with regard to this subject.

However, two images are available for public access and have been also uploaded in this repository. Both are ID card samples which present placeholder / invalid information and therefore do not belong to a specific person.

Run

Ensure that you've installed the client API for overfitted.io:

pip3 install git+https://github.com/overfitted-io/python-api-client.git

and placed your API key in the OVERFITTED_IO_API_KEY environment variable:

export OVERFITTED_IO_API_KEY=123456789abcdef

Then you can simply run the script:

python3 script.py

Expected Result

The expected result should correctly map the coordinates to the target image and extract the text as follows: