PRONUNCIATION VARIATION RULE EXTRACTION APPARATUS, PRONUNCIATION VARIATION RULE EXTRACTION METHOD, AND PRONUNCIATION VARIATION RULE EXTRACTION PROGRAM
1 Assignment
0 Petitions
Accused Products
Abstract
A problem to be solved is to robustly detect a pronunciation variation example and acquire a pronunciation variation rule having a high generalization property, with less effort. The problem can be solved by a pronunciation variation rule extraction apparatus including a speech data storage unit, a base form pronunciation storage unit, a sub word language model generation unit, a speech recognition unit, and a difference extraction unit. The speech data storage unit stores speech data. The base form pronunciation storage unit stores base form pronunciation data representing base form pronunciation of the speech data. The sub word language model generation unit generates a sub word language model from the base form pronunciation data. The speech recognition unit recognizes the speech data by using the sub word language model. The difference extraction unit extracts a difference between a recognition result outputted from the speech recognition unit and the base form pronunciation data by comparing the recognition result and the base form pronunciation data.
103 Citations
36 Claims
-
1-20. -20. (canceled)
-
21. A pronunciation variation rule extraction apparatus comprising:
-
a speech data storage unit for storing speech data; a surface form pronunciation storage unit for storing surface form pronunciation data representing surface form pronunciation of said speech data; a sub word language model generation unit for generating a sub word language model from said surface form pronunciation data; a speech recognition unit for recognizing said speech data by using said sub word language model; a difference extraction unit for extracting a difference between a recognition result outputted from said speech recognition unit and said surface form pronunciation data by comparing said recognition result and said surface form pronunciation data; and a language model weight control unit for controlling a weight value of said sub word language model, wherein said language model weight control unit outputs a plurality of weight values, said speech recognition unit recognizes said speech data for each of said plurality of weight values, and said language model weight control unit determines based on said difference at time when difference is extracted whether said weight value should be updated or not. - View Dependent Claims (22, 23, 24, 25, 26, 27)
-
-
28. A pronunciation variation rule extraction method comprising:
-
storing surface form pronunciation data representing surface form pronunciation of speech data; generating a sub word language model from said surface form pronunciation data; recognizing said speech data by using said sub word language model; extracting a difference between a recognition result of said recognizing and said surface form pronunciation data by comparing said recognition result and said surface form pronunciation data; and controlling a weight value of said sub word language model, wherein said controlling includes outputting a plurality of weight values, said recognizing includes recognizing said speech data for each of said plurality of weight values, and said controlling further includes determining based on said difference at time when said difference is extracted whether said weight value should be updated or not. - View Dependent Claims (29, 30, 31, 32)
-
-
33. A computer-readable recording medium which records a pronunciation variation rule extraction program which causes a computer to function as:
-
a speech data storage unit for storing speech data; a surface form pronunciation storage unit for storing surface form pronunciation data representing surface form pronunciation of said speech data; a sub word language model generation unit for generating a sub word language model from said surface form pronunciation data; a speech recognition unit for recognizing said speech data by using said sub word language model; a difference extraction unit for extracting a difference between a recognition result outputted from said speech recognition unit and said surface form pronunciation data by comparing said recognition result and said surface form pronunciation data; and a language model weight control unit for controlling a weight value of said sub word language model, wherein said language model weight control unit outputs a plurality of weight values, said speech recognition unit recognizes said speech data for each of said plurality of weight values, and said language model weight control unit determine based on said difference at a time when said difference is extracted whether said weight value should be updated or not. - View Dependent Claims (34, 35, 36)
-
Specification