| 摘要 |
PURPOSE:To provide a voice recognition device which can obtain a high sentence recognition rate as compared with a conventional example even when an inputted spoken voice includes a pause or redundant word. CONSTITUTION:The voice recognition device which performs voice recognition by calculating a probability value of a word to be recognized in the spoken voice sentence consisting of a character string inputted by referring to a specific statistical language model on the basis of one or plural words connected in front of the word sets the probability value of the language model of the pause in the spoken voice sentence or the word connected to the redundant word to 1 or a value close to 1, and calculates the probability of the language model of the work by skipping the pause or redundant word connected to a work other than the pause or redundant work across the pause or redundant work, thereby performing the voice recognizing process. |
