Method for creating a speech database for a target vocabulary in order to train a speech recognition system
First Claim
Patent Images
1. A method for creating a speech database for training of a speech recognition system with a target vocabulary, comprising:
- converting the words of the target vocabulary into a phonetic description so that the individual words are represented by a sequence of phonemes; and
concatenating, using a computer, segments of a generic spoken training text to form the words of the target vocabulary and thereby the speech database, each segment containing at least one phone, the segments being concatenated to form a sequence of phones which correspond respectively to the sequence of phonemes of the target vocabulary.
3 Assignments
0 Petitions
Accused Products
Abstract
The words of the target vocabulary are composed of segments, which have one or more phonemes, whereby the segments are derived from a training text that is independent from the target vocabulary. The training text can be an arbitrary generic text.
19 Citations
21 Claims
-
1. A method for creating a speech database for training of a speech recognition system with a target vocabulary, comprising:
-
converting the words of the target vocabulary into a phonetic description so that the individual words are represented by a sequence of phonemes; and concatenating, using a computer, segments of a generic spoken training text to form the words of the target vocabulary and thereby the speech database, each segment containing at least one phone, the segments being concatenated to form a sequence of phones which correspond respectively to the sequence of phonemes of the target vocabulary. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19)
-
-
20. A computer readable medium storing a program to control a computer to perform a method for creating a speech database for training of a speech recognition system with a target vocabulary, said method comprising:
-
converting the words of the target vocabulary into a phonetic description so that the individual words are represented by a sequence of phonemes, and concatenating segments of a generic spoken training text to form the words of the target vocabulary and thereby the speech database, each segment containing at least one phone, the segments being concatenated to form a sequence of phones which correspond respectively to the sequence of phonemes of the target vocabulary.
-
-
21. A method for creating a speech database for training of a speech recognition system with a target vocabulary, comprising:
-
converting the words of the target vocabulary into a phonetic description so that the individual words are represented by a sequence of phonemes; obtaining a generic spoken training text produced from speech of at least 100 speakers; and concatenating, using a computer, segments of the generic spoken training text to form the words of the target vocabulary and thereby the speech database, each segment containing at least one phone, the segments being concatenated to form a sequence of phones which correspond respectively to the sequence of phonemes of the target vocabulary, wherein the speech database is created without requiring speakers to speak the entire target vocabulary.
-
Specification