YOMITORIAYAMARIOSHIKIBETSUSHITEISEISURUHOHO

摘要

A method and system for identifying errors in recognition and for correcting such errors. During the recognition process, the most likely (or primary recognition) character and a second or alternative character (representing the next most likely recognition) together with a probability of the most likely character being correct. When an error is detected, as by the recognized characters on individual documents not totalling the total on a summary document, then characters with lowest probability in the correct place value are examined until the replacement of an alternate character for the most likely character allows the characters on the individual documents total the total on the summary document.