Electronic apparatus, speech recognition method thereof, and non-transitory computer readable recording medium

Patent metadata

Field	Value
Publication number	US-2018268808-A1
Application number	US-201615761506-A
Country	US
Kind code	A1
Filing date	Oct 21, 2016
Priority date	Oct 21, 2015
Publication date	Sep 20, 2018
Grant date	—

Abstract

An electronic apparatus is provided. The electronic apparatus according to an embodiment includes an audio input unit configured to receive sound sources from different positions and generate a plurality of voice signals, a pre-processor configured to perform pre-processing of the plurality of voice signals, and a voice recognition unit configured to perform voice recognition using the plurality of voice signals pre-processed by the pre-processor, and in response to a predetermined trigger being detected as a result of the voice recognition, generate trigger information, wherein the pre-processor is further configured to receive feedback on the trigger information generated by the voice recognition unit, change a pre-processing method according to the trigger information, process the plurality of voice signals using the changed pre-processing method, and generate enhanced voice signals.

First claim

What is claimed is:

1 . An electronic apparatus, comprising:
an audio input unit configured to receive sound sources from different positions and generate a plurality of voice signals; a pre-processor configured to perform pre-processing of the plurality of voice signals; and a voice recognition unit configured to perform voice recognition using the plurality of voice signals pre-processed by the pre-processor, and in response to a predetermined trigger being detected as a result of the voice recognition, generate trigger information, wherein the pre-processor is further configured to receive feedback on the trigger information generated by the voice recognition unit, change a pre-processing method according to the trigger information, process the plurality of voice signals using the changed pre-processing method, and generate enhanced voice signals.

2 . The electronic apparatus as claimed in claim 1 , further comprising:
a sensor configured to detect an utterance direction by using the plurality of voice signals, wherein the sensor is further configured to receive feedback on the trigger information generated by the voice recognition unit, change a search range and detect an utterance direction within the changed search range.

3 . The electronic apparatus as claimed in claim 2 , wherein the sensor is further configured to, if feedback on the trigger information is not received, detect an utterance direction with respect to all directions, and if feedback on the trigger information is received, detect an utterance direction only with respect to a search range corresponding to a pre-detected utterance direction.

4 . The electronic apparatus as claimed in claim 1 , further comprising:
a voice detector configured to distinguish between a voice interval and a voiceless interval by using the plurality of voice signals,
wherein the voice detector is further configured to receive feedback on the trigger information generated by the voice recognition unit, change a time determined as a voice interval after detecting a voice and distinguish between a voice interval and a voiceless interval based on the changed time determined as a voice interval.

5 . The electronic apparatus as claimed in claim 4 , wherein the voice detector is further configured to, if feedback on the trigger information is not received, distinguish between a voice interval and a voiceless interval by using a time pre-determined as a voice interval, and if feedback on the trigger information is received, distinguish between a voice interval and a voiceless interval by using a second time determined as a voice interval, which is longer than the time pre-determined as a voice interval.

6 . The electronic apparatus as claimed in claim 4 , wherein the pre-processor is further configured to change a pre-processing method for a plurality of voice signals based on the trigger information and whether a voice signal is a voice interval, and generate an enhanced voice signal using the changed pre-processing method.

7 . The electronic apparatus as claimed in claim 1 , wherein the pre-processor comprises:
a first element configured to generate a voiceless signal from which a voice is removed by using a plurality of voice signals; and a second element configured to generate an enhanced voice signal by using the generated voiceless signal, wherein the first element is configured to receive feedback on the trigger information generated by the voice recognition unit, change a degree of filtering and generate a voiceless signal.

8 . The electronic apparatus as claimed in claim 7 , wherein the second element is configured to receive feedback on the trigger information, change a degree of noise removal and generate an enhanced voice signal.

9 . The electronic apparatus as claimed in claim 1 , further comprising:
a display configured to display an image; and a controller configured to control the display to display an image corresponding to the voice recognition result.

10 . The electronic apparatus as claimed in claim 1 , further comprising:
a communicator configured to transmit the voice recognition result to an external device.

11 . The electronic apparatus as claimed in claim 1 , further comprising;
a speaker configured to output a voice output event corresponding to the voice recognition result.

12 . A method for voice recognition of an electronic apparatus, comprising:
receiving sound sources from different positions and generating a plurality of voice signals; performing pre-processing of the plurality of voice signals; performing voice recognition by using the pre-processed voice signals; in response to a predetermined trigger being detected as a result of the voice recognition, generating trigger information; receiving feedback on the generated trigger information and changing a method for performing the pre-processing according to the trigger information; and processing the plurality of voice signals using the changed pre-processing method and generating enhanced voice signals.

13 . The method as claimed in claim 12 , further comprising:
detecting an utterance direction by using the plurality of voice signals, wherein the changing comprises: receiving feedback on the generated trigger information and changing a search range; and detecting an utterance direction within the changed search range.

14 . The method as claimed in claim 13 , wherein the receiving of the feedback comprises, if feedback on the trigger information is not received, detecting an utterance direction with respect to all directions, and if feedback on the trigger information is received, detecting an utterance direction only with respect to a search range corresponding to a pre-detected utterance direction.

15 . The method as claimed in claim 12 , further comprising:
distinguishing between a voice interval and a voiceless interval by using the plurality of voice signals, wherein the changing comprises: receiving feedback on the generated trigger information and changing a time determined as a voice interval after detecting a voice; and distinguishing between a voice interval and a voiceless interval by using the changed time determined as a voice interval.

Assignees

• Samsung Electronics Co Ltd

Inventors

• Song Myung-Suk

Classifications

• G10L15/28

  Constructional details of speech recognition systems · CPC title

• G10L15/07 Primary

  to the speaker · CPC title

• G10L15/063

  Training · CPC title

• G10L2021/02166

  Microphone arrays; Beamforming · CPC title

• G10L21/0216

  characterised by the method used for estimating noise · CPC title