| 摘要 |
PURPOSE: An edition method is provided to display error characters or characters corresponding to voice and to present varied candidate words with good possibility via a morpheme analysis after character or voice recognition so that it enables a user to easily edit data. CONSTITUTION: The method comprises steps of performing a scanning operation(51), recognizing characters from scanned images(52), accessing a possibility of an error occurrence by analyzing morphemes of the recognized characters and replacing mistakable characters by characters with good possibility(53), separating the recognized characters by the unit of word, analyzing morphemes by the unit of word, extracting the pattern with the highest consistency by comparing the analyzed word morpheme with standard patterns classified by the unit of the word, and obtaining a recognition score(54), comparing the recognition score with a threshold value(55), in the case that Vi is less than the Vo, searching for similar words with a consistency degree more than a set value and displaying the searched words in other colour or font as candidates(56), and enabling a user to select one among the candidates by using a mouse(57).
|
