Speech recognition method for voice interaction, speech recognition programs and speech recognition system

摘要

Speech recognition while performing the dialogue without as well as effectively using a hard disk or limited memory, to produce an unnatural while even if there is a word that is not recognized or if the candidate is not found even after a certain period of time The object of the present invention is to provide a speech recognition method for the speech dialogue is performed, a voice recognition program and a speech recognition system, the step of recording in the storage speech recognition dictionary of one or more created for each scene it, (S1 step a), the step of inputting from the audio input unit to the speech of the speaker and (S4), performs speech recognition using a speech recognition dictionary for one or more audio data obtained by the word spotting voice of the speaker Create a response sentence based on a dialogue sequence if the (S7), the recognition is performed within a predetermined period of time, to facilitate re-enter the speaker when the recognition is not performed within a predetermined time step of creating the statement Kikikaeshi and (S8), composed and a (S9) step of speech synthesis statements Kikikaeshi response or statement has been created.