STATISTICAL LANGUAGE MODEL GENERATING DEVICE, SPEECH RECOGNITION DEVICE, STATISTICAL LANGUAGE MODEL GENERATING METHOD, SPEECH RECOGNIZING METHOD, AND PROGRAM
PROBLEM TO BE SOLVED: To make recognizable a natural utterance by a grammar system speech application program (hereinafter a "speech application") without time and labor nor cost as to a speech recognition device etc., which recognizes a natural utterance by using a grammar system speech a...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | PROBLEM TO BE SOLVED: To make recognizable a natural utterance by a grammar system speech application program (hereinafter a "speech application") without time and labor nor cost as to a speech recognition device etc., which recognizes a natural utterance by using a grammar system speech application. SOLUTION: The speech recognition device is equipped with a characteristic document generating means 14 for generating document data characteristic of an application from a grammar stored in a grammar storage part 12 and correction information stored in a correction information storage part 13 and storing the data in a characteristic document storage part 15, a topic shift means 16 for shifting a general topic stored in a general topic storage part 17 according to the document data and then storing it in a topic-after-shift storage part 18, a natural utterance recognizing means 21 for performing speech recognition by dictation of the topic after the shift and stores the recognition result in a recognition result storage part 22, a candidate sentence generating means 19 for generating a candidate sentence according to the grammar and storing it in a candidate sentence storage part 20, and a matching means 23 for matching the candidate sentence and the recognition result to determine a final recognition result. COPYRIGHT: (C)2005,JPO&NCIPI |
---|