Rendering followed by OCR and making sure that the computer doesn’t see more or less than the user does. Tricky and computionally more expensive.
Rendering followed by OCR and making sure that the computer doesn’t see more or less than the user does. Tricky and computionally more expensive.
And even then. I might question if what is rendered and then OCR is same as humans see on their screens... I am pretty sure there will be some tricks to change things enough for computer to get something different from humans.