Importing OCR pre-annotations for video annotation

I am trying to import OCR pre-annotations for video annotation. What is the format to be followed.