OCR for ID documents on video

preview_player
Показать описание
The goal of the project was to read, in short videos, the 3 OCR lines of the Spanish ID documents. Other countries have only 2 OCR lines and this lines appears also on passports (our software is valid for this other documents also with minimal customization). In this project we developed a full prototype using C++ and OpenCV on GNU/Linux. The prototype is fully functional (10 fps on an Intel Core i5). It makes one mistake on average per ID document (in our limited test cases of few IDs).
Рекомендации по теме
Комментарии
Автор

Estoy trabajando en un proyecto parecido. Te importaría mencionar los pasos seguidos para detectar el MRZ y separar los caracteres ? Es un proyecto de código abierto ? Gracias !

flaviusstefannicu
Автор

What is the technology you are using to read the text directly from the camera?

johnsonchua