Speech recognition using reject information

• US 10,304,449 B2
• Filed: 03/16/2016
• Issued: 05/28/2019
• Est. Priority Date: 03/27/2015
• Status: Active Grant

• Current Assignee: Panasonic Intellectual Property Management Co., Ltd. (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Intellectual Property Management Co., Ltd. (Panasonic Holdings Corporation)
• Inventors: Konuma, Tomohiro
• Primary Examiner(s): Pullias, Jesse S
• Application Number:
  US15/557,862
• Publication Number:
  US 20180053506A1
• Time in Patent Office: 1,168 Days
• Field of Search: 704231-257, 704270-275
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: H04N 21/42203 sound input device, e.g. microphone
• CPC Class Codes: H04N 21/42204 User interfaces specially adapted for controlling a client device through a remote control device; Remote control devices therefor constructive details of casings for the remote control device H01H9/0235; user interfaces for controlling a tuning device of a television receiver through a remote control H03J9/00; remote control of peripheral devices connected to a television receiver through the remote control device of the television receiver H04B1/205

Abstract

A voice recognition device includes a storage, a voice recognizer, and a reject information generator. The storage stores reject information for use in specifying a voice. The voice recognizer recognizes an input voice as a target word while defining, as one of conditions, a fact that a degree of likelihood between a voice specified by the reject information and the input voice is lower than a predetermined threshold value. In case where the voice recognizer has recognized the input voice as the target word in a predetermined state, the reject information generator determines the reject information, based on the input voice, and stores the reject information in the storage.

Claims

1. A voice recognition device that recognizes an input voice as a target word, the voice recognition device comprising:

a storage that stores reject information for use in specifying a voice;

a voice recognizer that recognizes the input voice as the target word at least when a degree of likelihood in an acoustic feature between the voice specified by the reject information and the input voice is lower than a predetermined threshold value;

a reject information generator that, in case where the voice recognizer has recognized a given input voice as a given target word based on the acoustic feature in a predetermined state, determines the given input voice as reject information for the given target word, and stores the reject information for the given target word in the storage; and

a controller that executes predetermined control processing in case where the voice recognizer has recognized that the given input voice is the given target word in a state that is not the predetermined state, and does not execute the control processing in case where the voice recognizer has recognized that the given input voice is the given target word in the predetermined state, wherein;

the reject information generator determines, as the reject information for the given target word, information which is related to a voice representing a word recognized without depending on the voice recognizer, based on the given input voice, and

the reject information generator acquires, from an external device, information which is related to a voice representing a word as a result of allowing the external device to perform voice recognition that is based on the given input voice, and determines, as the reject information, an information which is related to a voice representing the word.

2. The voice recognition device according to claim 1,

wherein the reject information generator stores the reject information for the given target word in the storage at a time of receiving a predetermined input in case where the voice recognizer has recognized that the given input voice is the given target word in the predetermined state.

3. The voice recognition device according to claim 1,

wherein:

the voice recognition device is provided in a television receiver including a microphone and a display,

the control processing is processing for supplying the display with power, and

the predetermined state is a state where the display is supplied with power.

4. The voice recognition device according to claim 1, further comprising

an output unit that outputs, to an outside, the reject information which is stored in the storage.

5. The voice recognition device according to claim 1,

wherein a state where input of the given target word is not awaited or a state where the given target word should not be recognized is defined as the predetermined state.

6. The voice recognition device according to claim 1, wherein

the voice recognition device is provided in a television receiver including a microphone and a display, and

a state where the display is supplied with power is defined as the predetermined state in case where the given target word is a word representing an instruction to supply the display with power.

7. A voice recognition system comprising:

a first voice recognition device that recognizes an input voice as a target word; and

a second voice recognition device that recognizes a word corresponding to the input voice,

wherein

the first voice recognition device includes;

a storage that stores reject information for use in specifying a voice;

a first voice recognizer that recognizes the input voice as the target word while defining, as one of conditions, a fact that a degree of likelihood between the voice specified by the reject information and the input voice is lower than a predetermined threshold value; and

a reject information generator that, in case where the first voice recognizer has recognized the input voice as the target word in a predetermined state, transmits the input voice to the second voice recognition device, receives, from the second voice recognition device, information which is related to a voice representing a word recognized by the second voice recognition device, based on the input voice, determines the received information as the reject information, and stores the reject information in the storage, and

the second voice recognition device has a second voice recognizer that recognizes a word, which corresponds to the input voice, based on the input voice received from the first voice recognition device, and that transmits the word to the first voice recognition device.

8. A voice recognition method for use in a voice recognition device including a storage that stores reject information for use in specifying a voice, and a voice recognizer that recognizes an input voice as a target word at least when a degree of likelihood in an acoustic feature between the voice specified by the reject information and the input voice is lower than a predetermined threshold value, the voice recognition method comprising:

determining whether or not the voice recognizer has recognized a given input voice as a given target word in a predetermined state where the given target word should not be recognized;

determining the given input voice as reject information for the given target word and storing the reject information for the given target word in the storage when it is determined that the voice recognizer has recognized that the given input voice is the given target word in the predetermined state; and

executing predetermined control processing in case where the voice recognizer has recognized that the given input voice is the given target word in a state that is not the predetermined state, and not executing the control processing in case where the voice recognizer has recognized that the given input voice is the given target word in the predetermined state, wherein;

information which is related to a voice representing a word recognized without depending on the voice recognizer is determined as the reject information for the given target word is determined, based on the given input voice, and

information which is related to a voice representing a word is acquired from an external device, as a result of allowing the external device to perform voice recognition that is based on the given input voice, and an information which is related to a voice representing the word is determined as the reject information.

9. A non-transitory storage medium storing a control program for allowing a processor of a voice recognition device to execute voice recognition processing, the voice recognition device including a storage that stores reject information for use in specifying a voice, and a voice recognizer that recognizes an input voice as a target word at least when a degree of likelihood in an acoustic feature between the voice specified by the reject information and the input voice is lower than a predetermined threshold value, the voice recognition processing including:

determining whether or not the voice recognizer has recognized a given input voice as a given target word in a predetermined state where the given target word should not be recognized;

determining the given input voice as the reject information for the given target word, and storing the reject information for the given target word in the storage in case where it is determined that the voice recognizer has recognized that the given input voice is the given target word in the predetermined state; and

executing predetermined control processing in case where the voice recognizer has recognized that the given input voice is the given target word in a state that is not the predetermined state, and not executing the control processing in case where the voice recognizer has recognized that the given input voice is the given target word in the predetermined state, wherein;

information which is related to a voice representing a word recognized without depending on the voice recognizer is determined as the reject information for the given target word is determined, based on the given input voice, and

information which is related to a voice representing a word is acquired from an external device, as a result of allowing the external device to perform voice recognition that is based on the given input voice, and an information which is related to a voice representing the word is determined as the reject information.

10. A voice recognition device that recognizes an input voice as a target word, the voice recognition device comprising:

a storage that stores reject information for use in specifying a voice;

a voice recognizer that recognizes the input voice as the target word while defining, as one of conditions, a fact that a degree of likelihood between the voice specified by the reject information and the input voice is lower than a predetermined threshold value;

a reject information generator that, in case where the voice recognizer has recognized the input voice as the target word in a predetermined state, determines the reject information, based on the input voice, and stores the reject information in the storage; and

a controller that executes predetermined control processing in case where the voice recognizer has recognized that the input voice is the target word in a state that is not the predetermined state, and does not execute the control processing in case where the voice recognizer has recognized that the input voice is the target word in the predetermined state, wherein

the reject information generator determines, as the reject information, information which is related to a voice representing a word recognized without depending on the voice recognizer, based on the input voice, and

the reject information generator acquires, from an external device, information which is related to a voice representing a word as a result of allowing the external device to perform voice recognition that is based on the input voice, and determines, as the reject information, an information which is related to a voice representing the word.

11. A voice recognition method for use in a voice recognition device including a storage that stores reject information for use in specifying a voice, and a voice recognizer that recognizes an input voice as a target word while defining, as one of conditions, a fact that a degree of likelihood between the voice specified by the reject information and the input voice is lower than a predetermined threshold value, the voice recognition method comprising:

determining whether or not the voice recognizer has recognized the input voice as the target word in a predetermined state where the target word should not be recognized;

determining the reject information based on the input voice and storing the reject information in the storage when it is determined that the voice recognizer has recognized that the input voice is the target word in the predetermined state;

executing predetermined control processing in case where the voice recognizer has recognized that the input voice is the target word in a state that is not the predetermined state, and not executing the control processing in case where the voice recognizer has recognized that the input voice is the target word in the predetermined state;

determining, as the reject information, information which is related to a voice representing a word recognized without depending on the voice recognizer, based on the input voice; and

acquiring, from an external device, information which is related to a voice representing a word as a result of allowing the external device to perform voice recognition that is based on the input voice, and determining, as the reject information, an information which is related to a voice representing the word.

12. A non-transitory storage medium storing a control program for allowing a processor of a voice recognition device to execute voice recognition processing, the voice recognition device including a storage that stores reject information for use in specifying a voice, and a voice recognizer that recognizes an input voice as a target word while defining, as one of conditions, a fact that a degree of likelihood between the voice specified by the reject information and the input voice is lower than a predetermined threshold value, the voice recognition processing including:

determining whether or not the voice recognizer has recognized the input voice as the target word in a predetermined state where the target word should not be recognized;

determining the reject information, based on the input voice and storing the reject information in the storage in case where it is determined that the voice recognizer has recognized that the input voice is the target word in the predetermined state;

executing predetermined control processing in case where the voice recognizer has recognized that the input voice is the target word in a state that is not the predetermined state, and not executing the control processing in case where the voice recognizer has recognized that the input voice is the target word in the predetermined state;

determining, as the reject information, information which is related to a voice representing a word recognized without depending on the voice recognizer, based on the input voice; and

acquiring, from an external device, information which is related to a voice representing a word as a result of allowing the external device to perform voice recognition that is based on the input voice, and determining, as the reject information, an information which is related to a voice representing the word.