Assume the returned text contains coordinates (x, y), you can extract tables from the image, horizontal and vertial lines, then you know coordinates of each cell, and it's easy to put text to the correct cell.