Speech recognition method and apparatus

• US 7,565,290 B2
• Filed: 06/24/2005
• Issued: 07/21/2009
• Est. Priority Date: 06/29/2004
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Kuboyama, Hideo, Komori, Yasuhiro, Fukada, Toshiaki
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): SAINT CYR, LEONARD
• Application Number:
  US11/165,167
• Publication Number:
  US 20050288929A1
• Time in Patent Office: 1,488 Days
• Field of Search: 704/256, 704/238, 704/243, 704/245
• US Class Current: 704/251
• CPC Class Codes: G10L 15/142 Hidden Markov Models [HMMs]

Abstract

A speech recognition apparatus includes a word dictionary having recognition target words, a first acoustic model which expresses a reference pattern of a speech unit by one or more states, a second acoustic model which is lower in precision than said first acoustic model, selection means for selecting one of said first acoustic model and said second acoustic model on the basis of a parameter associated with a state of interest, and likelihood calculation means for calculating a likelihood of an acoustic feature parameter with respect to said acoustic model selected by said selection means.

Claims

1. A speech recognition apparatus comprising:

(a) a memory for storing;

(i) a word dictionary having recognition target words, said word dictionary comprising a tree structure in which the recognition target words share a predetermined speech unit;

(ii) a first acoustic model which expresses a reference pattern of the speech unit by one or more states; and

(iii) a second acoustic model which is lower in precision than said first acoustic model;

(b) selection means for selecting a state of interest from the tree structure;

(c) checking means for checking the number of branches of the selected state; and

(d) likelihood calculation means for calculating a likelihood of an acoustic feature parameter for states immediately succeeding the selected state using the first acoustic model, if the number of branches of the selected state is equal to or more than a predetermined value, and otherwise calculating a likelihood of an acoustic feature parameter for states immediately succeeding the selected state using the second acoustic model;

wherein in calculating a likelihood with respect to a state of interest by using said second acoustic model, if likelihood calculation using said first acoustic model has been performed for a state having the same speech unit alignment as that of the state of interest, said likelihood calculation means reuses a result of the likelihood calculation as a result of likelihood calculation for the state of interest.

2. The apparatus according to claim 1, wherein with respect to states for which said selection means has selected said second acoustic model, said likelihood calculation means calculates likelihoods by using said second acoustic model and re-calculates likelihoods by using said first acoustic model with respect to only states which exhibit a likelihood higher than a threshold.

3. The apparatus according to claim 1, wherein with respect to states for which said selection means has selected said second acoustic model, said likelihood calculation means calculates likelihoods by using said second acoustic model and re-calculates a likelihood by using said first acoustic model with respect to only a predetermined number of states of the states which are selected in decreasing order of likelihood.

4. The apparatus according to claim 1, wherein the number of mixed distributions which express said second acoustic model is smaller than the number of mixed distributions which express said first acoustic model.

5. The apparatus according to claim 1, wherein

said first acoustic model is a triphone model, and

said second acoustic model is a monophone model.

6. A speech recognition method of performing speech recognition in a speech recognition apparatus by using (i) a word dictionary having recognition target words, the word dictionary comprising a tree structure in which the recognition target words share a predetermined speech unit, (ii) a first acoustic model which expresses a reference pattern of the speech unit by one or more states, and (iii) a second acoustic model which is lower in precision than the first acoustic model, the method comprising the steps of:

selecting a state of interest from the tree structure;

checking the number of branches of the selected state; and

calculating a likelihood of an acoustic feature parameter for states immediately succeeding the selected state using the first acoustic model, if the number of branches of the selected state is equal to or more than a predetermined value, and otherwise calculating a likelihood of an acoustic feature parameter for states immediately succeeding the selected state using the second acoustic model;

wherein in calculating a likelihood with respect to a state of interest by using said second acoustic model, if likelihood calculation using said first acoustic model has been performed for a state having the same speech unit alignment as that of the state of interest, said likelihood calculation means reuses a result of the likelihood calculation as a result of likelihood calculation for the state of interest.

7. A program for implementing a speech recognition method defined in claim 6 by using a computer.

8. A computer-readable storage medium storing executable instructions for causing a computer to perform the method of claim 6.