| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a text correction device which uses and consecutively corrects voice recognition results, and outputs the correction in a text within a prescribed limited delay time independent of a length of an utterance unit or the like. <P>SOLUTION: A text section estimation unit determines whether or not a length of recognition result word string data written in a recognition result word string data storage unit is equal to or longer than a preliminarily determined prescribed value N and, upon determining the length to be equal to or longer than the prescribed value N, performs matching processing of text word string data stored in a text word string data storage unit and the recognition result word string data to calculate a similarity between the text word string data and the recognition result word string data and selects the text word string data as a correction candidate on the basis of the calculated similarity. An error correction unit corrects a section which does not match the text word string data selected by the text section estimation unit on the basis of the same. <P>COPYRIGHT: (C)2012,JPO&INPIT |
