| 摘要 |
PROBLEM TO BE SOLVED: To satisfactorily prevent it that erroneous recognition is to be generated by voices even when a user speaks voices of 'er-' or the like unconsciously. SOLUTION: In an S20 continuous to a well-known registering and confirming processing (S10), this voice recognizing method speaks a question that what is this device which displays a map on a display ? from a speaker and when an answer with respect to this speaking is inputted with voices (S30: YES), it executes a voice recognizing processing (S40). In this device, answers such as 'a navi', 'a navigation', 'a navigation system', 'I can't understand', 'I don't know' are preliminarily prepared with respect to the question and in a succeeding S50, voices recognized in the S40 are compared with those of the prepared answers. Then, voice recognition is executed by excluding the effect of unconscious expression of the 'er-' or the like by extracting voices other than those of expression used in the above described answers as the uncocious speech (S60) and by storing them (S70).
|
