How to tell OCR to Read Specific Region... can anyone help?

Hello All,

I am using custom vision to identify regions in some photos. I have it trained and it is doing a great job generating the proper tags. These tags have bounding areas in which are some text that I would like to extract.... I would like to use OCR to extract that text, but I would like to send it the specific region to analyze. Has anyone done this or something like it?

Thanks in advance for any help.