To provide a pronunciation evaluating device and a program capable of evaluating pronunciation coping with special cases such as a voiceless section, phonemic insertion, phonemic replacement, and phonemic lack.
The pronunciation evaluating device comprises: an input speech data acquisition section storing one or more teacher data which is speech data of an object of comparison and which is based on phonological HMM, receiving input of speeches, dividing the received speech into frames, and acquiring the input speech data for each divided frame; a special speech detecting section which detects, based on the input speech data for each frame, input of special speeches; an evaluating section for evaluating, based on the teacher data and input speech data and the detection result of the special speech detection section, the received speech; and an output section which outputs the evaluation result of the evaluating section. Thus, the pronunciation evaluating device can detect occurrence of special phenomena such as a voiceless section, phonemic replacement, phonemic insert, and phonemic lack, and can evaluate the speech with high accuracy.
COPYRIGHT: (C)2006,JPO&NCIPI
Hideyuki Watanabe
Yamada Reiko
JP2001265211A | ||||
JP9258765A | ||||
JP6110494A | ||||
JP2003345380A | ||||
JP7503559A | ||||
JP2004117530A |