VOICE PHRASE RECOGNITION METHOD AND VOICE RECOGNITION DEVICE

摘要

PROBLEM TO BE SOLVED: To improve the precision of a recognition result without interactions with a user. SOLUTION: A category of a supplied voice phrase is determined by key sub-phrase spotting processing; and when the category is defined, limited vocabulary corresponding to the defined category is used to execute second recognition processing and a second recognition result is generated. First recognition processing is executed on the basis of comprehensive vocabulary, and key sub- phrase spotting processing is executed on the basis of the first recognition result.