Greetings of the day,
I would like to introduce myself as a deep learning engineer who have a experience of 2 years in this field. I'm good with images and had worked in object detection, segmentation, classification as well as GAN's. I have been developing OCR related projects for financial institutions like banks and insurance agencies for assisting their onboarding needs. I had also developed a auto key-value pair Recognition engine for invoices.
As for the project you mentioned, there are many possible python based platforms (like tesseract, craft etc.,) that could be used to complete the project on time and if run time is our priority, we can use image processing approaches to solve this which can run at considerable FPS.
we can take this conversation further, if you are still open with this project.
best regards