JP2007286174A - Electronic apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an electronic apparatus capable of surely performing voice recognition by selecting a sound model most suitable for each user. <P>SOLUTION: In a digital television receiver 100, a sound model pattern corresponding to identification information extracted by an identification information extracting means (a CPU121, an identification information extraction program 123c) is acquired by a sound model pattern acquiring means (the CPU 121, a sound model pattern acquiring means 123d). The voice recognition of voice information acquired by a voice information acquiring section 11 is performed by using the acquired sound model pattern, and command information is extracted by a command information extracting means (the CPU121, a command information extracting program 123e), and control is performed by a control means based on the extracted command information. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an electronic apparatus capable of performing various controls by a voice recognition function.

  In recent years, as electronic devices such as televisions and videos have become more multifunctional and sophisticated, it has become necessary to support various functions and enable simple operation. Electronic devices using recognition technology are used.

  Electronic devices using voice recognition technology include, for example, a facsimile machine with an answering machine function, and when voice is input as a normal operation aid, voice recognition is performed based on a predetermined number of words, When performing speech recognition of a specific speaker such as ID verification, there is a device that registers speech data of a phrase to be used in advance and performs speech recognition (see Patent Document 1).

  It also has a voice sample-equipment number correspondence table that stores device numbers and corresponding phonetic notation in tabular form, and searches the voice sample-equipment number correspondence table using speech recognition results to detect corresponding device numbers. By doing so, there is an apparatus for specifying a controlled device (see Patent Document 2).

  In addition, the controller device generates a multiplexed signal obtained by multiplexing the input voice and the command to the electronic device and transmits the multiplexed signal to the electronic device. The electronic device separates the voice and the command from the received multiplexed signal. There is a device that recognizes a separated voice and controls its own operation in accordance with the recognized content and the separated command (see Patent Document 3).

Also, a small-sized portable remote control device having a voice recognition function, by inputting a command by voice, a command corresponding to the voice command is wirelessly transmitted to a device to be controlled, and the device There is a device that controls the above (see Patent Document 4).
In general, the speech recognition technology can be divided into two types: specific speaker speech recognition and non-specific speaker speech recognition. The specific speaker voice recognition technology is a technology that performs voice recognition by creating an acoustic model from a voice of a user's individual based on the premise that the user is a specific individual. For this reason, since a precise acoustic model only for the user can be created, a high recognition rate can be obtained. On the other hand, the unspecified speaker speech recognition technology is a technology for performing speech recognition by creating a general acoustic model in advance based on the premise that the user is unspecified. For this reason, although the recognition performance is inferior to the specific speaker voice recognition, there is a feature that voice recognition that anyone can use is provided from the beginning.
JP 2001-309080 A JP 2003-330483 A JP 2003-250061 A JP 2003-114694 A

  However, the acoustic model varies greatly depending on the use of speakers, children, adults, etc., the usage environment such as in a telephone or a car, and in the electronic device using the voice recognition function disclosed in Patent Documents 1 to 4, each user Since the acoustic model relating to the frequency pattern of the voice signal that is most suitable for this is not selected, voice recognition may not be performed correctly.

  An object of the present invention is to provide an electronic device that can select an optimal acoustic model for each user and perform voice recognition more reliably.

In order to solve the above-described problem, the invention according to claim 1 is an electronic device that performs voice recognition of voice information input from the outside and is controlled by command information obtained based on the voice recognition result.
Identification information composed of audio information for identifying a user and an acoustic model pattern related to the frequency pattern of the audio signal can be stored in association with each other, and a plurality of acoustic model patterns can be associated with one identification information. Acoustic model pattern storage means configured to be capable of storage;
Command information storage means for storing command information including voice information for controlling the electronic device;
Audio information acquisition means for acquiring audio information;
Identification information extraction means for performing voice recognition of the voice information acquired by the voice information acquisition means, and extracting identification information stored in the acoustic model pattern storage means;
Acoustic model pattern acquisition means for acquiring an acoustic model pattern corresponding to the identification information extracted by the identification information extraction means from the acoustic model pattern storage means;
Command information for performing voice recognition of the voice information acquired by the voice information acquisition means using the acoustic model pattern acquired by the acoustic model pattern acquisition means and extracting command information stored in the command information storage means Extraction means;
Control means for performing control based on the command information extracted by the command information extraction means;
Calculating means for calculating a speech recognition rate by the command information extracting means for each identification information and for each acoustic model pattern;
Voice recognition rate storage means for storing the voice recognition rate calculated by the calculation means for each identification information and for each acoustic model pattern;
Display control means for causing the display means to display the control content by the control means corresponding to the command information extracted by the command information extraction means;
Specifying means for specifying whether to execute or stop the control content displayed on the display means by the display control means,
When the acoustic model pattern acquisition unit stores a plurality of acoustic model patterns in association with the identification information extracted by the identification information extraction unit and stores them in the acoustic model pattern storage unit, the acoustic model pattern acquisition unit stores Acquired acoustic model patterns in descending order of the speech recognition rate of the stored acoustic model pattern associated with the identification information,
The control means performs control corresponding to the command information when the designation means designates execution of the control content.

The invention according to claim 2 is an electronic device that performs voice recognition of voice information input from the outside and is controlled by command information obtained based on a voice recognition result.
Acoustic model pattern storage means for storing identification information composed of voice information for identifying a user and an acoustic model pattern related to a frequency pattern of the voice signal in association with each other;
Command information storage means for storing command information including voice information for controlling the electronic device;
Audio information acquisition means for acquiring audio information;
Identification information extraction means for performing voice recognition of the voice information acquired by the voice information acquisition means, and extracting identification information stored in the acoustic model pattern storage means;
Acoustic model pattern acquisition means for acquiring an acoustic model pattern corresponding to the identification information extracted by the identification information extraction means from the acoustic model pattern storage means;
Command information for performing voice recognition of the voice information acquired by the voice information acquisition means using the acoustic model pattern acquired by the acoustic model pattern acquisition means and extracting command information stored in the command information storage means Extraction means;
Control means for performing control based on the command information extracted by the command information extraction means.

The invention according to claim 3 is the invention according to claim 2,
The acoustic model pattern storage means is configured to be capable of storing a plurality of acoustic model patterns in association with one identification information,
The acoustic model pattern acquisition unit is associated with the identification information when a plurality of acoustic model patterns are associated with the identification information extracted by the identification information extraction unit and stored in the acoustic model pattern storage unit. If the command information cannot be extracted as a result of performing voice recognition of the voice information by the command information extraction means, the other information associated with the identification information is obtained. The acoustic model pattern is acquired.

The invention according to claim 4 is the invention according to claim 3,
Calculating means for calculating a speech recognition rate by the command information extracting means for each identification information and for each acoustic model pattern;
Voice recognition rate storage means for storing the voice recognition rate calculated by the calculation means for each identification information and for each acoustic model pattern;
When the acoustic model pattern acquisition unit stores a plurality of acoustic model patterns in association with the identification information extracted by the identification information extraction unit and stores them in the acoustic model pattern storage unit, the acoustic model pattern acquisition unit stores The acoustic model patterns are acquired in descending order of the speech recognition rate of the stored acoustic model pattern associated with the identification information.

The invention according to claim 5 is the invention according to any one of claims 2 to 4,
Display control means for causing the display means to display the control content by the control means corresponding to the command information extracted by the command information extraction means;
Further comprising designation means for designating whether to execute or stop the control content displayed on the display means by the display control means,
The control means performs control corresponding to the command information when the designation means designates execution of the control content.

According to the first aspect of the invention, the identification information extraction unit can perform voice recognition of the voice information acquired by the voice information acquisition unit and extract the identification information stored in the acoustic model pattern storage unit. The voice recognition rate by the command information extraction unit can be calculated for each identification information and for each acoustic model pattern by the calculation unit, and a plurality of acoustic model patterns are included in the identification information extracted by the identification information extraction unit. Is stored in the acoustic model pattern storage means, the acoustic model pattern acquisition means stores the voice recognition rate of the acoustic model pattern associated with the identification information stored in the voice recognition rate storage means. The acoustic model pattern can be acquired in descending order, and the command information extracting means can acquire the acoustic model pattern. Using the obtained acoustic model pattern by, it performs speech recognition of the speech information obtained by the speech information acquisition unit, it is possible to extract the command information stored in the command information storage means.
Therefore, when the electronic device acquires voice information, the user who has input the voice information can be identified, and an acoustic model pattern optimal for the user can be acquired and voice recognition can be performed. Recognition can be performed.
Further, the display control means can cause the display means to display the control content by the control means corresponding to the command information extracted by the command information extraction means, and the control means displayed on the display means by the display control means by the designation means. Whether the content is to be executed or canceled can be specified. When the control means specifies that the control content is to be executed by the specifying means, control corresponding to the command information can be performed.
Therefore, before the control based on the voice recognition result is performed, the control content can be displayed and confirmed to the user, and malfunctions can be suitably reduced.

According to the second aspect of the present invention, the identification information extraction unit can perform voice recognition of the voice information acquired by the voice information acquisition unit and extract the identification information stored in the acoustic model pattern storage unit. The acoustic model pattern acquisition unit can acquire the acoustic model pattern corresponding to the identification information extracted by the identification information extraction unit from the acoustic model pattern storage unit, and the command information extraction unit acquires the acoustic model pattern by the acoustic model pattern acquisition unit. The voice information acquired by the voice information acquisition means can be recognized using the acoustic model pattern, and the command information stored in the command information storage means can be extracted, and the command information extraction means can be extracted by the control means. Control can be performed based on the command information extracted by the above.
Therefore, when the electronic device acquires voice information, the user who has input the voice information can be identified, an acoustic model pattern suitable for the user can be acquired, and voice recognition can be performed more reliably. Voice recognition can be performed.

According to the invention described in claim 3, it is needless to say that the same effect as that of the invention described in claim 2 can be obtained, and the acoustic model pattern storage means has a plurality of acoustic model patterns for one piece of identification information. Are stored in association with each other, and when a plurality of acoustic model patterns are associated with the identification information extracted by the identification information extraction means and stored in the acoustic model pattern storage means, the acoustic model pattern acquisition means If one acoustic model pattern associated with the identification information is obtained and the command information extraction means performs speech recognition of the speech information, the command information cannot be extracted. Another obtained acoustic model pattern can be acquired.
Therefore, the user can select any acoustic model pattern from a plurality of acoustic model patterns and perform speech recognition. For example, when the frequency pattern changes due to a change in the user's physical condition, the user can use one acoustic model pattern. Even if the voice recognition fails, another acoustic model pattern can be selected, so that the voice recognition can be suitably performed.

According to the invention described in claim 4, it is a matter of course that the same effect as that of the invention described in claim 3 can be obtained, and the calculation means executes the command for each identification information and for each acoustic model pattern. When the speech recognition rate by the information extraction unit can be calculated and a plurality of acoustic model patterns are associated with the identification information extracted by the identification information extraction unit and stored in the acoustic model pattern storage unit, the acoustic model pattern The acquisition unit can acquire the acoustic model patterns in descending order of the speech recognition rate of the acoustic model pattern associated with the identification information stored in the speech recognition rate storage unit.
Therefore, it is possible to select the optimal acoustic model pattern for each user based on the speech recognition rate and perform speech recognition more reliably.

According to the invention described in claim 5, it is needless to say that the same effect as the invention described in any one of claims 2 to 4 can be obtained, and it is extracted by the command information extracting means by the display control means. The control content by the control means corresponding to the command information can be displayed on the display means, and the designation means designates whether the control content displayed on the display means is executed or stopped by the display control means When the control means specifies that the control content is to be executed, the control means can perform control corresponding to the command information.
Therefore, before the control based on the voice recognition result is performed, the control content can be displayed and confirmed to the user, and malfunctions can be suitably reduced.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
Note that in this embodiment, a digital television receiver is described as an example of the electronic device. However, the electronic device is not limited to this, and may be a home appliance such as a video or an air conditioner.

First, with reference to FIG. 1, the configuration of the entire device and the configuration of the main part will be described.
The digital television receiver 100 of this embodiment receives, for example, a television broadcast signal (hereinafter referred to as a broadcast signal), converts the received broadcast signal into a predetermined output signal, and outputs video / audio data. A device main body unit 1 and a remote control device with a microphone (hereinafter referred to as a remote controller with a microphone) 2 as a voice information output means that allows a user to output voice information to the device main body unit 1 wirelessly. Has been.

Next, the configuration of the main part of the digital television receiver 100 will be described.
The device main unit 1 includes an antenna 3 that receives a broadcast signal, a tuner 4 that selects a broadcast signal of a predetermined broadcast channel from the broadcast signal received by the antenna 3, demodulation of the broadcast signal output from the tuner 4, and Each of the demodulator 5 that performs error correction processing, the descramble 6 that cancels the scramble signal for preventing illegal viewing added to the broadcast signal demodulated by the demodulator 5, and the broadcast signal that is canceled by the descramble 6 By separating and extracting the data, a demultiplexer 7 for acquiring video / audio data, a decoder 8 for performing expansion processing on the video / audio data acquired by the demultiplexer 7, and OSD data for the video data expanded by the decoder 8 An OSD (On-Screen Character Display) processing unit 9 to be added, a decoder 8 and a table for outputting video / audio data processed by the OSD processing unit 9 The image receiving unit 10 as a means, an audio information acquisition unit 11 that acquires audio information, a control unit 12 that performs overall control of the entire device main body 1, and a control bus 13 that connects the units are configured. .

The voice information acquisition unit 11 uses, for example, a microphone, and acquires voice as it is as voice information.
Note that the audio information acquisition unit 11 is not limited to acquiring audio as audio information as it is, and may be designed to acquire the audio signal as audio information after the audio is converted into an audio signal.

  For example, as shown in FIG. 1, the control unit 12 includes a CPU (Central Processing Unit) 121, a RAM (Random Access Memory) 122, a storage unit 123, and the like.

  The CPU 121 performs various control operations according to the acoustic model pattern table, command table, and various processing programs stored in the storage unit 123.

  The RAM 122 includes a program storage area for expanding a processing program executed by the CPU 121, a data storage area for storing input data, a processing result generated when the processing program is executed, and the like.

The storage unit 123 stores a system program that can be executed by the device main unit 1, various processing programs that can be executed by the system program, data that is used when these various processing programs are executed, and processing results that are calculated by the CPU 121. Store data etc. Note that the program is stored in the storage unit 123 in the form of a computer-readable program code.
Specifically, in the storage unit 123, for example, as shown in FIG. 1, an acoustic model pattern table 123a, a command table 123b, an identification information extraction program 123c, an acoustic model pattern acquisition program 123d, a command information extraction program 123e, and a control A program 123f, a calculation program 123g, a display control program 123h, and the like are stored.

The acoustic model pattern table 123a can store identification information including audio information for identifying a user and an acoustic model pattern related to a frequency pattern of the audio signal in association with each other, and a plurality of identification information can be stored with respect to one identification information. The table is configured to be able to store an acoustic model pattern in association with each other, and stores a speech recognition rate calculated by a calculation program 123g described later for each identification information and for each acoustic model pattern.
Specifically, in the acoustic model pattern table 123a, for example, as shown in FIG. 2, the user ID-1 has identification information composed of the audio information “Ichiro”, and a plurality of acoustics are associated with the identification information. The model patterns A, B, and C are stored in association with the speech recognition rate for each acoustic model pattern.
The storage unit 123 functions as an acoustic model pattern storage unit and a voice recognition rate storage unit by storing the acoustic model pattern table 123a.

The command table 123 b is a table that stores command information including voice information for controlling the device main body 1.
The storage unit 123 functions as a command information storage unit by storing the command table 123b.

The identification information extraction program 123c is a program that causes the CPU 121 to realize a function of performing voice recognition based on the voice information acquired by the voice information acquisition unit 11 and extracting the identification information.
Here, speech recognition is realized by, for example, using Julius speech recognition software and accumulating enormous speech patterns and language pattern statistical data. The basic principle of speech recognition is to maintain the frequency pattern of an acoustic model (phoneme (approximately equivalent to one Roman character) or syllable (equivalent to one kana character)) and refer to the word dictionary while referring to the word dictionary. It performs matching with the syllable string signal obtained by cutting out a single syllable part from the signal and recognizes the voice.
Specifically, the CPU 121 executes voice recognition of voice information acquired by the voice information acquisition unit 11 by executing the identification information extraction program 123c, using a preset acoustic model pattern, and an acoustic model pattern table Matching identification information is extracted from the identification information stored in 123a.
The CPU 121 functions as an identification information extraction unit by executing the identification information extraction program 123c.

The acoustic model pattern acquisition program 123d is a program that causes the CPU 121 to realize a function of acquiring, from the acoustic model pattern table 123a, an acoustic model pattern corresponding to the identification information extracted by executing the identification information extraction program 123c.
Specifically, the CPU 121 acquires an acoustic model pattern corresponding to the identification information extracted by the execution of the identification information extraction program 123c from the acoustic model pattern table 123a by executing the acoustic model pattern acquisition program 123d. When a plurality of acoustic model patterns are associated with the identification information extracted by the identification information extraction program 123c and stored in the acoustic model pattern table 123a, the voice recognition rate of the acoustic model pattern associated with the identification information The acoustic model pattern is acquired in descending order. For example, as shown in FIG. 2, when the identification information consisting of “Ichiro” is extracted, the CPU 121 first acquires the acoustic model pattern A having the highest speech recognition rate, and then acquires the acoustic model pattern B, Acquired in the order of acoustic model pattern C.
The CPU 121 functions as an acoustic model pattern acquisition unit by executing the acoustic model pattern acquisition program 123d.

The command information extraction program 123e is a program that causes the CPU 121 to realize voice recognition of voice information acquired by the voice information acquisition unit 11 and to extract command information based on the voice recognition result.
Specifically, by executing the command information extraction program 123e, the CPU 121 performs voice recognition of the voice information using the acoustic model pattern acquired by the acoustic model pattern acquisition program 123d, and is stored in the command table 123b. Extracts matching command information from the command information.
The CPU 121 functions as a command information extraction unit by executing the command information extraction program 123e.

The control program 123f is a program that causes the CPU 121 to realize a function of controlling the device main body 1 based on command information based on the voice recognition result by executing the control program 123f.
Specifically, the CPU 121 displays the control content “channel up” corresponding to the command information extracted by the execution of the command information extraction program 123e on the image receiving unit 10 by executing the display control program 123h described later. When the execution of the displayed control content is instructed by the remote controller 2 with the microphone, the control corresponding to the command information is performed by executing the control program 123f.
The CPU 121 functions as a control unit by executing the control program 123f.

The calculation program 123g is a program that causes the CPU 121 to realize a function of calculating a voice recognition rate based on a voice recognition result.
Specifically, by executing the calculation program 123g, the CPU 121 executes the command information extraction program 123e for each piece of identification information stored in the acoustic model pattern table 123a and for each acoustic model pattern. The recognition rate is calculated. The calculation method uses, for example, a value obtained by dividing the number of successful extractions by the number of executions of the command information extraction program 123e by executing the command information extraction program 123e.
The CPU 121 functions as a calculation unit by executing the calculation program 123g.

The display control program 123h is a program that causes the CPU 121 to realize a function of causing the image receiving unit 10 to display control contents corresponding to command information based on the voice recognition result.
Specifically, as shown in FIG. 3, the CPU 121 executes the display control program 123 h to thereby display the control content corresponding to the command information extracted by the execution of the command information extraction program 123 e as the display unit. 10 is displayed.
The CPU 121 functions as a display control unit by executing the display control program 123h.

Next, the operation of the device main unit 1 according to the present invention will be described with reference to FIG. 4 using a digital television receiver as an embodiment of the present invention as an example. The explanation will be made assuming that Ichiro, who is a user, wants to channel up.
First, in step S1, the audio information acquisition unit 11 acquires “Ichiro” and “Channel Up” that are audio information. Next, in step S2, the CPU 121 performs voice recognition of the voice information “Ichiro” acquired by the voice information acquisition unit 11 by executing the identification information extraction program 123c, and the identification stored in the acoustic model pattern table 123a. Extract information "Ichiro".

  Next, when the identification information can be extracted in step S3 (step S3; Yes), the process proceeds to step S4. On the other hand, if the identification information cannot be extracted in step S3 (step S3; No), the process returns to step S1.

  Next, in step S4, the CPU 121 acquires an acoustic model pattern corresponding to the identification information from the acoustic model pattern table 123a by executing the acoustic model pattern acquisition program 123d. In such a case, as shown in FIG. 2, a plurality of acoustic model patterns A, B, and C are stored in association with the identification information “Ichiro”, but the acoustic model pattern A having the highest speech recognition rate is stored. To get.

  Next, in step S5, the CPU 121 executes the command information extraction program 123e to perform voice recognition of the voice information “channel up” acquired by the voice information acquisition unit 11 using the acquired acoustic model pattern. The command information stored in the command table 123b is extracted.

  Next, in step S6, when command information can be extracted by executing the command information extraction program 123e (step S6; Yes), the process proceeds to step S7. On the other hand, when command information cannot be extracted in step S6 (step S6; No), the process returns to step S4, and the CPU 121 acquires the acoustic model pattern again by executing the acoustic model pattern acquisition program 123d. . At this time, when a plurality of acoustic model patterns are associated with the identification information and stored in the acoustic model pattern table 123a, the CPU 121 has a high speech recognition rate of the acoustic model pattern by executing the acoustic model pattern acquisition program 123d. An acoustic model pattern is acquired in order. In such a case, the acoustic model pattern B having the highest speech recognition rate next to the acoustic model pattern A is acquired.

  Next, in step S <b> 7, the CPU 121 displays the control content “channel up” corresponding to the command information extracted by the execution of the command information extraction program 123 e on the image receiving unit 10 by executing the display control program 123 h.

  Next, when the control content displayed on the image receiving unit 10 is executed in step S8 (step S8; Yes), the selection of “Yes” display is selected as shown in FIG. Designate and proceed to step S9. On the other hand, when the selection designation of the display of “No” is performed (step S8; No), this process is terminated.

  Next, in step S9, the CPU 121 performs control corresponding to the command information by executing the control program 123f, and ends this processing.

According to the digital television receiver 100 according to the present invention described above, the CPU 121 executes voice recognition of the voice information acquired by the voice information acquisition unit 11 by executing the identification information extraction program 123c, and the acoustic model pattern. The identification information stored in the table 123a can be extracted, and by executing the calculation program 123g, the speech recognition rate by the execution of the command information extraction program 123e is determined for each identification information and for each acoustic model pattern. A plurality of acoustic model patterns are associated with the identification information extracted by the execution of the identification information extraction program 123c and stored in the acoustic model pattern table 123a by the execution of the acoustic model pattern acquisition program 123d. If the acoustic model The acoustic model pattern can be acquired in descending order of the speech recognition rate of the acoustic model pattern associated with the identification information stored in the turntable 123a. By executing the command information extraction program 123e, the acoustic model pattern acquisition program Using the acoustic model pattern acquired by executing 123d, the voice information acquired by the voice information acquisition unit 11 can be recognized, and the command information stored in the command table 123b can be extracted.
Therefore, when the electronic device acquires voice information, it is possible to identify a user who has input the voice information, acquire an acoustic model pattern optimal for the user, and perform voice recognition.
In addition, the CPU 121 can display the control content corresponding to the command information extracted by the execution of the command information extraction program 123e on the image receiving unit 10 by executing the display control program 123h. By executing the display control program 123h, it is possible to specify whether to execute or cancel the control content displayed on the image receiving unit 10, and by executing the control program 123f, the control content is controlled by the remote controller with microphone 2. When it is designated to execute, control corresponding to the command information can be performed.
Therefore, before the control based on the voice recognition result is performed, the control content can be displayed and confirmed to the user, and malfunctions can be suitably reduced.

The present invention is not limited to the above embodiment, and various improvements and design changes may be made without departing from the spirit of the present invention.
For example, the designating means is not limited to the remote controller with a microphone according to the present invention, and may be designed so that it can be designated by an operation key on an operation panel provided in the apparatus main body.
The acoustic model pattern acquisition unit may be designed such that an arbitrary acoustic model pattern can be manually acquired by a user.
The calculation of the speech recognition rate is not limited to whether or not the command information is extracted by the command information extracting means, but the command information is extracted and the control content corresponding to the command information displayed on the display means is specified by the specifying means. The design may be such that whether or not execution is designated can be added to the calculation criteria.

It is a block diagram which shows the principal part structure of the digital television receiver concerning this invention. It is a figure which shows an example of the acoustic model pattern table in this invention. It is a figure which shows the example of a display by the display control means in this invention. It is a flowchart which shows the operation | movement process of the apparatus main body part in this invention.

Explanation of symbols

100 Digital television receiver (electronic equipment)
1 Device body 2 Remote control with microphone (designating means)
10 Image receiver (display means)
11 Voice information acquisition unit (voice information acquisition means)
121 CPU (identification information extraction means, acoustic model pattern acquisition means, command information extraction means, control means, calculation means, display control means)
123a Acoustic model pattern table (acoustic model pattern storage means, speech recognition rate storage means)
123b Command table (command information storage means)
123c Identification information extraction program (identification information extraction means)
123d Acoustic model pattern acquisition program (acoustic model pattern acquisition means)
123e Command information extraction program (command information extraction means)
123f Control program (control means)
123g Calculation program (calculation means)
123h Display control program (display control means)

Claims (5)

In an electronic device that performs voice recognition of voice information input from the outside and is controlled by command information obtained based on the voice recognition result,
Identification information composed of audio information for identifying a user and an acoustic model pattern related to the frequency pattern of the audio signal can be stored in association with each other, and a plurality of acoustic model patterns can be associated with one identification information. Acoustic model pattern storage means configured to be capable of storage;
Command information storage means for storing command information including voice information for controlling the electronic device;
Audio information acquisition means for acquiring audio information;
Identification information extraction means for performing voice recognition of the voice information acquired by the voice information acquisition means, and extracting identification information stored in the acoustic model pattern storage means;
Acoustic model pattern acquisition means for acquiring an acoustic model pattern corresponding to the identification information extracted by the identification information extraction means from the acoustic model pattern storage means;
Command information for performing voice recognition of the voice information acquired by the voice information acquisition means using the acoustic model pattern acquired by the acoustic model pattern acquisition means and extracting command information stored in the command information storage means Extraction means;
Control means for performing control based on the command information extracted by the command information extraction means;
Calculating means for calculating a speech recognition rate by the command information extracting means for each identification information and for each acoustic model pattern;
Voice recognition rate storage means for storing the voice recognition rate calculated by the calculation means for each identification information and for each acoustic model pattern;
Display control means for causing the display means to display the control content by the control means corresponding to the command information extracted by the command information extraction means;
Specifying means for specifying whether to execute or stop the control content displayed on the display means by the display control means,
When the acoustic model pattern acquisition unit stores a plurality of acoustic model patterns in association with the identification information extracted by the identification information extraction unit and stores them in the acoustic model pattern storage unit, the acoustic model pattern acquisition unit stores Acquired acoustic model patterns in descending order of the speech recognition rate of the stored acoustic model pattern associated with the identification information,
The electronic device according to claim 1, wherein the control means performs control corresponding to the command information when the designation means designates execution of the control content. In an electronic device that performs voice recognition of voice information input from the outside and is controlled by command information obtained based on the voice recognition result,
Acoustic model pattern storage means for storing identification information composed of voice information for identifying a user and an acoustic model pattern related to a frequency pattern of the voice signal in association with each other;
Command information storage means for storing command information including voice information for controlling the electronic device;
Audio information acquisition means for acquiring audio information;
Identification information extraction means for performing voice recognition of the voice information acquired by the voice information acquisition means, and extracting identification information stored in the acoustic model pattern storage means;
Acoustic model pattern acquisition means for acquiring an acoustic model pattern corresponding to the identification information extracted by the identification information extraction means from the acoustic model pattern storage means;
Command information for performing voice recognition of the voice information acquired by the voice information acquisition means using the acoustic model pattern acquired by the acoustic model pattern acquisition means and extracting command information stored in the command information storage means Extraction means;
Control means for performing control based on the command information extracted by the command information extraction means;
An electronic device comprising: The acoustic model pattern storage means is configured to be capable of storing a plurality of acoustic model patterns in association with one identification information,
The acoustic model pattern acquisition unit is associated with the identification information when a plurality of acoustic model patterns are associated with the identification information extracted by the identification information extraction unit and stored in the acoustic model pattern storage unit. If the command information cannot be extracted as a result of performing voice recognition of the voice information by the command information extraction means, the other information associated with the identification information is obtained. The electronic device according to claim 2, wherein an acoustic model pattern is acquired. Calculating means for calculating a speech recognition rate by the command information extracting means for each identification information and for each acoustic model pattern;
Voice recognition rate storage means for storing the voice recognition rate calculated by the calculation means for each identification information and for each acoustic model pattern;
When the acoustic model pattern acquisition unit stores a plurality of acoustic model patterns in association with the identification information extracted by the identification information extraction unit and stores them in the acoustic model pattern storage unit, the acoustic model pattern acquisition unit stores The electronic device according to claim 3, wherein the acoustic model patterns are acquired in descending order of the speech recognition rate of the stored acoustic model pattern associated with the identification information. Display control means for causing the display means to display the control content by the control means corresponding to the command information extracted by the command information extraction means;
Further comprising designation means for designating whether to execute or stop the control content displayed on the display means by the display control means,
5. The electronic apparatus according to claim 2, wherein the control unit performs control corresponding to the command information when the specification unit specifies that the control content is to be executed. 6. .

Images