METHOD OF RECOGNIZING SPEECH AND ELECTRONIC DEVICE THEREOF

• US 20140019131A1
• Filed: 07/12/2013
• Published: 01/16/2014
• Est. Priority Date: 07/13/2012
• Status: Abandoned Application

• Current Assignee: Korea University Research and Business Foundation (Korea University), Samsung Electronics Co. Ltd.
• Original Assignee: Korea University Research and Business Foundation (Korea University), Samsung Electronics Co. Ltd.
• Inventors: LEE, Jae-won, YOOK, Dong-suk, LIM, Hyeon-taek, KIM, Tae-yoon
• Application Number:
  US13/940,848
• Publication Number:
  US 20140019131A1
• Time in Patent Office: Days
• US Class Current: 704/254
• CPC Class Codes: G10L 15/05 Word boundary detection
• CPC Class Codes: G10L 15/142 Hidden Markov Models [HMMs]
• CPC Class Codes: G10L 15/18 using natural language modelling

Abstract

A method of recognizing a speech and an electronic device thereof are provided. The method includes: segmenting a speech signal into a plurality of sections at preset time intervals; performing a phoneme recognition with respect to one of the plurality of sections of the speech signal by using a first acoustic model; extracting a candidate word of the one of the plurality of sections of the speech signal by using the phoneme recognition result; and performing a speech recognition with respect to the one the plurality of sections the speech signal by using the candidate word.

Claims

1. A method of recognizing speech in an electronic device, the method comprising:

segmenting a speech signal into a plurality of sections at preset time intervals;

performing a phoneme recognition with respect to one of the plurality of sections of the speech signal based on a first acoustic model;

extracting a candidate word of the one of the plurality of sections of the speech signal based on a result of the phoneme recognition; and

performing speech recognition with respect to the one of the plurality of sections of the speech signal based on the candidate word.

2. The method of claim 1, wherein the performing of the phoneme recognition further comprises:

deleting at least one last phoneme of a plurality of phonemes of the one of the plurality of sections of the speech signal based on a segmented viterbi algorithm,

wherein the at least one deleted phoneme is used to perform a phoneme recognition with respect to a next section of the speech signal following the one the plurality of sections.

3. The method of claim 1, wherein the extracting comprises:

extracting a similar phoneme pronounced similarly to the recognized phoneme; and

generating a word graph for extracting the candidate word of the one of the plurality of sections based on the recognized phoneme and the similar phoneme.

4. The method of claim 3, wherein the performing the speech recognition comprises:

calculating a Gaussian probability of the speech signal of the one of the plurality of sections based on a second acoustic model; and

outputting a word string having a highest probability in the word graph based on the second acoustic model and a language model.

5. The method of claim 4, wherein the first and second acoustic models are different from each other.

6. The method of claim 1, wherein the performing the phoneme recognition, the extracting the candidate word, and the performing the speech recognition are performed in parallel by different cores of the electronic device.

7. An electronic device comprising:

a speech signal input part configured to receive a speech signal;

a speech signal segmenter configured to segment the speech signal input through the speech signal input part into a plurality of sections at preset time intervals;

a phoneme recognizer configured to perform a phoneme recognition with respect to one of the plurality of sections of the speech signal based on a first acoustic model;

a candidate word extractor configured to extract a candidate word of the one of the plurality of sections of the speech signal based on a result of the phoneme recognition; and

a speech recognizer configured to perform speech recognition with respect to the one of the plurality of sections of the speech signal based on the candidate word.

8. The electronic device of claim 7, wherein the phoneme recognizer is configured to delete at least one last phoneme of a plurality of phonemes of the one of the plurality of sections of the speech signal based on a segmented viterbi algorithm to perform the phoneme recognition,

wherein the at least one deleted phoneme is used to perform a phoneme recognition with respect to a next section of the speech signal following the one of the plurality of sections.

9. The electronic device of claim 7, wherein the candidate word extractor is configured to extract a similar phoneme pronounced similarly to the recognized phoneme and generate a word graph for extracting a candidate word of the one of the plurality of sections based on the recognized phoneme and the similar phoneme.

10. The electronic device of claim 9, wherein the speech recognizer is configured to calculate a Gaussian probability of the speech signal of the one of the plurality of sections based on a second acoustic model and output a word string having a highest probability in the word graph based on the second acoustic model and a language model to perform the speech recognition.

11. The electronic device of claim 10, wherein the first acoustic model of the phoneme recognizer and the second acoustic model of the speech recognizer are different from each other.

12. The electronic device of claim 7, wherein the phoneme recognizer, the candidate word extractor, and the speech recognizer are realized as different cores of the electronic device.

13. A method of recognizing speech in an electronic device, the method comprising:

receiving a speech signal;

segmenting the received speech signal into a plurality of sections;

performing phoneme recognition on a first section of the plurality of sections at a first time;

performing phoneme recognition on a second section of the plurality of sections, and extracting a candidate word of the first section at a second time; and

performing phoneme recognition on a third section of the plurality of sections, extracting a candidate word of the second section, and performing speech recognition on the first section of the plurality of sections at a third time.

14. The method of claim 13, wherein a phoneme recognition operation, a candidate word extracting operation based on a recognized phoneme, and speech recognition based on the candidate word are performed in parallel.

15. The method of claim 14, wherein the performance of the phoneme recognition, the extracting operation, and the speech recognition operation are performed through different cores.