Hello, we are using the Real-Time Recognition SDK to read data from Italian ID like Italian health insurance card or Driving licence, we think the cause may be the background of the documents that make the text hard to read.

I have attached an example of health insurance card. Any advice to improve your data?