VOICE RECOGNITION APPARATUS AND METHOD FOR PERFORMING VOICE RECOGNITION

• US 20090210227A1
• Filed: 02/12/2009
• Published: 08/20/2009
• Est. Priority Date: 02/15/2008
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Suzuki, Kaoru, Sugiyama, Hiroshi, Koga, Toshiyuki, Yamamoto, Daisuke
• Granted Patent: US 8,155,968 B2
• Time in Patent Office: Days
• US Class Current: 704/246
• CPC Class Codes: G10L 15/20 Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech G10L21/02 takes precedence
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command

Abstract

A voice recognition apparatus includes: a voice recognition module that performs a voice recognition for an audio signal during a voice period; a distance measurement module that measures a current distance between the user and an voice input module; a calculation module that calculates a recommended distance range, in which being estimated that an S/N ratio exceeds a first threshold, based on the voice characteristic; and a display module that displays the recommended distance range and the current distance.

Claims

1. A voice recognition apparatus comprising:

an audio input module that receives an audio input and outputs an audio signal, the audio input module having a gain being configured to be adjustable;

a voice recognition module that detects a voice period, where a voice activity by a user is detected, in the audio signal and performs a voice recognition for the audio signal during the voice period;

a first level measurement module that measures a voice level of the audio signal in the voice period and outputs the voice level;

a second level measurement module that measures a noise level of the audio signal in a noise period and outputs the noise level, the noise period being a time period except the voice period;

a first calculation module that calculates an S/N ratio that is a ratio of the voice level to the noise level;

a distance measurement module that measures a current distance between the user and the voice input module;

a first memory module that stores a first threshold corresponding to an S/N ratio at which the voice recognition module is capable to perform the voice recognition with a given recognition rate;

a second memory module that stores a voice characteristic having a set of the voice level, the current distance and the gain, by which the voice recognition module has succeeded in performing the voice recognition;

a second calculation module that calculates a recommended distance range for the current distance, in which being estimated that the S/N ratio exceeds the first threshold, based on the voice characteristic; and

a display module that displays the recommended distance range and the current distance.

2. The apparatus according to claim 1 further comprising a third memory module that stores a second threshold corresponding to an upper limit of the audio signal permissible to be output from the audio input module,

wherein the second calculation module calculates the recommended distance range by calculating a distance range, in which being estimated that the voice level unexceeds the second threshold, based on the voice characteristic and the gain and setting the calculated distance range as the recommended distance range.

3. The apparatus according to claim 2 further comprising:

a fourth memory module that stores a third threshold corresponding to an upper limit of the noise level permissible in the audio signal; and

a control module that controls the gain of the voice recognition module to control the noise level to unexceed the third threshold.

4. The apparatus according to claim 3 further comprising:

a determination module that determines whether the current distance is within the recommended distance range; and

a notification module that notifies to the user a result of the determination by the determination module.

5. The apparatus according to claim 1, wherein the second memory module stores the voice characteristic for each of a plurality of users with an identification information that is unique to each of the users.

6. The apparatus according to claim 5, wherein the second calculation module calculates the recommended distance range based on the voice characteristic specified by a given identification information.

7. The apparatus according to claim 6 further comprising a user identification module that identifies the user by whom the voice is made and obtains the identification information corresponding to the identified user,

wherein the second calculation module calculates the recommended distance range based on the voice characteristic specified by the identification information obtained by the user identification module.

8. A method for performing a voice recognition, the method comprising:

receiving an audio input to output an audio signal with a gain being configured to be adjustable;

detecting a voice period, where a voice activity by a user is detected, in the audio signal to perform the voice recognition for the audio signal during the voice period;

measuring a voice level of the audio signal in the voice period to output the voice level;

measuring a noise level of the audio signal in a noise period to output the noise level, the noise period being a time period except the voice period;

calculating an S/N ratio that is a ratio of the voice level to the noise level;

measuring a current distance from the user;

storing a first threshold corresponding to an S/N ratio at which the voice recognition is capable to be performed to have a given recognition rate;

storing a voice characteristic having a set of the voice level, the current distance and the gain, by which the voice recognition has been succeeded;

calculating a recommended distance range for the current distance, in which being estimated that the S/N ratio exceeds the first threshold, based on the voice characteristic; and

displaying the recommended distance range and the current distance.