PHONEME ERROR ACQUISITION DEVICE, DICTIONARY ADDITION DEVICE, SPEECH RECOGNITION DEVICE, PHONEME ERROR ACQUISITION METHOD, SPEECH RECOGNITION METHOD, AND PROGRAM

摘要

PROBLEM TO BE SOLVED: To improve speech recognition precision on non-native-speaker's utterance.;SOLUTION: A non-native phoneme frequency totaling part 20 totals a phoneme error frequency of each phoneme pair by comparing a non-native phoneme recognition result of phoneme recognition on a speech of non-native utterance with a non-native phoneme transcription obtained by transcribing a phoneme of the non-native utterance. A native phoneme frequency totaling part 21 totals a phoneme error frequency of each phoneme pair by comparing a native phoneme recognition result of phoneme recognition on a speech of native utterance with a native phoneme transcription obtained by transcribing a phoneme of the native utterance. A phoneme error acquisition part 22 generates information representing a degree of likelihood of an error of the non-native utterance from the native utterance for each phoneme pair by comparing the phoneme error frequency of the non-native utterance with the phoneme error frequency of the native utterance, and acquires a phoneme error obtained by adding the information to a phoneme error.;SELECTED DRAWING: Figure 3;COPYRIGHT: (C)2018,JPO&INPIT