Lex processing is Transym’s unique technique for correcting common OCR mistakes. It does not use standard dictionaries; it uses character frequency/probability tables. This allows TOCR to cope with new words and common misspellings. Setting Lex on improves accuracy by using the context of the character and those around it. TOCR 3.3 for the first time allows TOCR to automatically decide whether to use Lex when processing an image.