Could an ocr reader do a first pass on the easy characters to guess the font, eliminating the problem of random font for the second pass of parsing more difficu

Could an ocr reader do a first pass on the easy characters to guess the font, eliminating the problem of random font for the second pass of parsing more difficult characters.