JP2001005480A - User uttering discriminating device and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To instruct a user that the pronunciation of the user causes malrecognition of his voice by comparing a recognized character string and a set character string and discriminating whether these strings are matched with each other or not. SOLUTION: When a pronunciation correction mode is selected, a control circuit 2 reads a character string having a specific title from an external memory 7, displays the string on a display device 8 and also displays a screen to urge a user for uttering. At the same time, the circuit 2 synthesizes the voice corresponding to the read character string and outputs the voice from a speaker 9. Thus, the user is able to listen to a standard uttering. When the user utters the character string having the specific title, his voice is inputted into the circuit 2 through a microphone 17 and transmitted to a voice processing section 10. The section 10 conducts a voice recognition process of the voice uttered by the user to recognize the title. The circuit 2 compares the uttered character string and the set character string. If no agreement is reached, a message is displayed on the device 8 to indicate the fact that the words are wrong.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to various devices having a voice recognition function (for example, a car navigation system).
And a recording medium.

[0002]

2. Description of the Related Art In a car navigation system having a voice recognition function, for example, when a place name is input to display a map of the place name or input an operation command,
The user can input by uttering a place name or an operation command. Therefore, a troublesome key operation is not required, so that the system is very convenient for the user.

[0003]

[0007] In the case of the above configuration, words uttered by a user may be erroneously recognized.
In that case, the user may see a map with a place name they do not want, or there may be no such map,
An operation (action) of a command not intended by the user was sometimes performed. One of the causes of the erroneous recognition of the voice is, for example, a case where the pronunciation of the user is not a standard pronunciation (for example, a case where the user is accentuated or the pronunciation is unclear).

[0004] However, in the case of the above-mentioned conventional configuration, even if a false recognition is uttered due to a user's pronunciation,
The map is configured such that a map of a place name that the user does not desire is displayed, or that the map does not exist, but the cause of the misrecognition is completely unknown. For this reason, even if the pronunciation of the user has a cause, the user often thinks that the performance of the voice recognition function of the car navigation system is poor.

Accordingly, an object of the present invention is to provide a user pronunciation determination apparatus and a recording medium that can teach a user to cause a user to misunderstand a voice when the user has a cause in pronunciation. It is in.

[0006]

According to the first aspect of the present invention, there is provided a speech recognition apparatus for recognizing a character string corresponding to the set speech by performing a speech recognition process on a speech when the user reads and utters a set character string. Means, and display means for displaying a character string recognized by the voice recognition means, and
A configuration is provided that includes a determination unit that compares the recognized character string with the set character string to determine whether they match. According to this configuration, when there is a cause in the pronunciation of the user and the voice is erroneously recognized, the user can easily recognize that there is a problem in his / her pronunciation.

According to the second aspect of the present invention, the determination means teaches the difference between the recognized character string and the set character string.
Unrecognized characters can be clearly known. Therefore,
The user can correct the pronunciation of an unrecognized character.

According to the third aspect of the present invention, since the determination unit teaches the degree of coincidence between the recognized character string and the set character string, the user can set the recognized character string as the recognized character string. It is possible to clearly know the degree of coincidence of the character string, and to use it for correction of pronunciation.

According to the fourth aspect of the present invention, the user repeatedly performs the process of speaking the set character string until the recognized character string matches the set character string. Can be reliably corrected, and erroneous recognition of voice can be further prevented.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the present invention is applied to a car navigation system will be described below with reference to the drawings. First, FIG. 2 is a block diagram illustrating the overall schematic configuration of the car navigation system 1 according to the present embodiment. As shown in FIG. 2, the car navigation system 1 includes a control circuit 2, a position detector 3, a map data input device 4, an operation switch group 5, a communication device 6, an external memory 7, a display device (Display means) 8, speaker 9,
An audio processing unit 10 is provided.

The control circuit 2 has a function of controlling the overall operation of the car navigation system 1, and is configured as a normal computer (for example, a microcomputer). That is, the control circuit 2 includes a CPU, a ROM, an R
It is configured to include an AM, an I / O, and a bus (not shown) for connecting these. The control circuit 2 has functions as a character string setting unit, a determination unit, and a repetition unit of the present invention.

The position detector 3 is a GPS (Global P
The ositioning system includes a receiver 11, a gyroscope 12, a distance sensor 13, and a geomagnetic sensor 14. The position detector 3 is configured to detect the current position of the vehicle while interpolating each other by the four sensors 11 to 14, and has a highly accurate position detection function. When the position detection accuracy is not so required, the position detector 3 may be configured by any one of the four sensors 11 to 14. Further, the position detector 3 may be configured by combining a steering rotation sensor, a sensor of each rolling wheel, and the like.

The map data input device 4 is, for example, a DVD-R
It is configured by a reading device that reads a recording medium such as an OM, and is a device for inputting map data, map matching data, landmark data, HTML information (Internet information), and the like. The recording medium may be, for example, a CD-ROM or a memory card.

The display device 8 is composed of, for example, a liquid crystal display or the like, and is capable of color display and a display screen 15 (FIG. 4) capable of clearly displaying maps, characters, images, and the like.
To FIG. 7). The display screen 15 of the display device 8 is configured so that a current position mark of the vehicle, map data, and additional data such as a guide route to be displayed on the map can be displayed in a superimposed manner.

The operation switch group 5 includes a touch switch (touch panel) provided on the upper surface of the display screen 15 of the display device 8 and a mechanical push switch provided on the periphery of the display screen 15. I have. The communication device 6 is, for example, a device that transmits and receives infrastructure data, and has a VICS (Vehicle Information & Communic).
ation System) device. Further, the speaker 9 provides voices for guidance of various operation procedures and route guidance, and the like.
Also, it is a device for outputting a standard voice (generation) of a character string for making the user utter.

The car navigation system 1
The control circuit 2 has a function of automatically setting an optimal route (guidance route) from the current position to the destination when a user operates the operation switch group 5 or the remote controller 18 to set a destination. To perform a map matching process for positioning a map on a map, and general information such as HT
A function for browsing ML information (Internet information) is provided. As a method for automatically setting the optimum route, for example, the Dijkstra method is known.

The voice processing unit 10 includes a voice recognition unit 16 and a microphone 1 for inputting voice uttered by a user.
7 are provided. The above speech recognition unit 1
6 is a dictionary unit 18a and a collating unit 18 as shown in FIG.
b, and a storage unit 19 for storing learning results based on past matching and external situations. The voice processing unit 10 has a function as voice recognition means of the present invention.

The external situation stored in the storage unit 19 includes information on the current location of the vehicle notified from the control circuit 2, information on the display state of the display device 8, and the recognition rules set by the user. (For example, a reserved word to be recognized as a command when recognizing a voice). The dictionary unit 18a stores a recognition target vocabulary including a word registered by the user and a structure representing the relation between the vocabulary.

When there is a voice input via the microphone 17, the recognition unit 18b selects a character (character string) closest to the voice input from the dictionary unit 18a based on the past learning result. Is configured to be performed. Note that this collation processing is executed using a known program (algorithm) for collation processing. Further, the recognizing unit 18b compares the result of the collation with the reserved word stored in the storage unit 19, and outputs to the control circuit 2 as a corresponding command if they match, or as kana character data if they do not match. Is configured to be executed.

Next, the operation of the above configuration, specifically, the operation in the pronunciation correction mode (pronunciation correction operation) used to correct the voice (pronunciation) uttered by the user will be described with reference to FIGS. 7 will also be described. The flowchart of FIG. 1 shows the contents of the partial control corresponding to the above-mentioned pronunciation correcting mode in the control program stored in the control circuit 2.

First, on the menu selection screen (not shown) of the car navigation system 1, for example, the pronunciation correction mode is selected by operating a selection switch or inputting a selection command by voice. When this selection is made, the pronunciation correcting mode is started, and the process proceeds to "YES" in step S100, and the control circuit 2 executes a process of reading a character string having a specific name from the external memory 7 (step S100). S110). The character string of this specific name is a character string for pronunciation correction, that is, a character string for causing the user to utter (a character string set in advance). In the present embodiment, “Aichiken Toyoshi” is used. Was. It should be noted that another character string may be used as the character string having the specific name, or may be appropriately selected from a plurality of character strings, or may be selected by the user. It may be configured as follows.

Then, the process proceeds to step S120, where the control circuit 2 displays the read character string on the display device 8 in a display form as shown in FIG. 4, and displays a screen for urging the user to speak. At the same time, the process proceeds to step S130, where the control circuit 2 synthesizes a voice corresponding to the read character string, and outputs (generates) the voice from the speaker 9.
The process of synthesizing the voice is performed, for example, by the voice processing unit 10.
It is configured to be executed by. Then, the user can hear a standard utterance (pronunciation) of the character string having the specific name by listening to the sound output from the speaker 9.

Here, it waits for the user to read the character string having the specific name and to utter (pronounce) it (step S14).
0). In this standby state, when the user reads the character string having the specific name and utters the voice, the voice is input to the control circuit 2 through the microphone 17 and sent to the voice processing unit 10. Then, the process proceeds to “YES” in step S140,
The process proceeds to step S150, in which the voice uttered by the user is subjected to voice recognition processing in the voice processing unit 10, thereby recognizing (identifying) a character string corresponding to the voice, that is, a name. Then, the character string of this specific result (ie,
The uttered character string) is sent to the control circuit 2.

Subsequently, the process proceeds to step S160, where the control circuit 2 performs a process of comparing the uttered character string with the set character string (that is, a character string having a specific name). Here, if they match, the process proceeds to “match” in step S160, and the process proceeds to step S180.
The fact that the voice uttered by the user matches, that is, that the voice uttered by the user is standard and has been correctly recognized is displayed on the display device 8. In this case, as shown in FIG. 5, it is configured to display the character string of the recognition result and to display a message of "meeting", and the displayed screen is the utterance matching screen.
Then, after that, the above-described pronunciation correction mode is ended.

On the other hand, if it is determined in step S160 that the uttered character string does not match the set character string, the process proceeds to “mismatch” and proceeds to step S170.
In this step S170, as shown in FIG.
A result of the disagreement, that is, a message indicating that the words are different, and a character string of the recognition result (uttered character string) are displayed on the display device 8. Here, a character different from the set character string in the character string of the recognition result is configured to be displayed in reverse video, for example. Further, a message indicating that the language is different may be output as a voice to teach the user.

In order to find the different characters, if both characters have the same number, the characters may be compared one by one. If the numbers of characters are different, an appropriate algorithm is used to execute a process to find as many matching portions (character strings) as possible for both character strings, thereby obtaining different characters and missing characters. It is preferable to display (point out) differences such as. Further, the degree of coincidence between the recognized character string and the set character string (for example, in the case of FIG. 6, eight of nine characters match, the degree of coincidence is 8/9, that is, about 89% May be displayed on the display screen. still,
Instead of displaying the degree of coincidence, it may be configured to output it by voice and to teach the user.

Then, as shown in FIG. 6, after displaying the recognition result, for example, for a predetermined time (time of about several seconds to several tens of seconds), the process proceeds to step S120, where the user is prompted as shown in FIG. On the other hand, it is configured to display a message prompting utterance again (that is, an utterance prompting screen). Subsequently, the processing after step S130 is repeated as described above. Note that the processing in step S130, that is, the processing in which the car navigation system 1 utters the set character string with standard pronunciation may be omitted in the second and subsequent times, or may be repeated. It may be configured as follows.

In this embodiment, the recognition result is displayed for a predetermined time in step S170. Instead, the recognition result is displayed and whether or not to continue the pronunciation correction mode is determined. It may be configured to ask the user. In the case of this configuration, when the user makes a response to continue the pronunciation correction mode, the process may proceed to step S120. Further, when the user responds to end the pronunciation correction mode, the pronunciation correction mode may be terminated.

In this embodiment having the above-described structure, the voice when the user reads and utters the set character string is subjected to voice recognition processing to recognize the character string corresponding to the voice, and then the voice is processed. The recognized character string is displayed, and the recognized character string is compared with the set character string to determine whether or not they match. According to this configuration, when there is a cause in the user's pronunciation and the voice is erroneously recognized, the user can easily recognize that there is a problem in his / her pronunciation.

Further, in the above embodiment, as shown in FIG. 6, the difference between the recognized character string and the set character string is displayed (taught). The unrecognized character can be clearly known, and the pronunciation of the unrecognized character can be corrected so that it becomes a standard pronunciation. Further, in the above embodiment, the configuration is such that the degree of coincidence between the recognized character string and the set character string is displayed (taught), so that the user can check the degree of coincidence between the recognized character string and the set character string. Can be clearly understood.

Further, in the above-described embodiment, the user repeatedly performs the process of uttering the set character string until the recognized character string matches the set character string. The pronunciation can be reliably corrected, and erroneous recognition of the voice can be further prevented.

In the above embodiment, the pronunciation correction mode is selected and executed by operating a selection switch on the menu selection screen of the car navigation system 1, for example. However, the present invention is not limited to this. Absent. For example, by operating a specific operation switch or operating a plurality of specific operation switches on a screen displaying a map or another navigation screen, the pronunciation correction mode can be selected and executed. It may be configured as follows.

In the above embodiment, a program for operating the car navigation system 1 (ie, a program for realizing each function as a character string setting unit, a judgment unit, a repetition unit, and a voice recognition unit) is controlled by a control circuit. 2 was stored in the ROM. In the case of this configuration, the ROM storing the program may be configured to be replaceable for the printed circuit board configuring the control circuit 2. In addition, the above program is stored on a CD-ROM or D-ROM.
A configuration in which the program is stored in a recording medium such as a VD-ROM, a rewritable nonvolatile memory such as a flash memory is provided in the control circuit 2, and the program is transferred from the recording medium to the nonvolatile memory. You may.

In the above embodiment, the user pronunciation determination device of the present invention is applied to the car navigation system 1. However, the present invention is not limited to this, and may be applied to a portable navigation system. Alternatively, the present invention may be applied to various electric devices having a voice recognition function.

[Brief description of the drawings]

FIG. 1 is a flowchart showing an embodiment of the present invention.

FIG. 2 is a block diagram of a car navigation system.

FIG. 3 is a block diagram of a speech recognition unit.

FIG. 4 is a diagram showing a voice prompting screen.

FIG. 5 is a diagram showing an utterance matching screen;

FIG. 6 is a diagram showing an utterance mismatch screen;

FIG. 7 is a diagram showing the second and subsequent vocalization prompting screens;

[Explanation of symbols]

1 is a car navigation system, 2 is a control circuit (character string setting means, determination means and repetition means), 4 is a map data input device, 5 is an operation switch group, 8 is a display device, 9 is a speaker, and 10 is a voice processing unit. (Voice recognition means), 15 is a display screen, 16 is a voice recognition unit, 17 is a microphone, 1
Reference numeral 8 denotes a recognition unit, 18a denotes a dictionary unit, 18b denotes a collation unit, and 19 denotes a storage unit.

Claims (5)

[Claims] 1. A character string setting means for setting a character string to be uttered by a user, and a character corresponding to the voice by performing a voice recognition process on a voice when the user reads and utters the set character string. Voice recognition means for recognizing a string, display means for displaying a character string recognized by the voice recognition means, and comparing the recognized character string with the set character string to determine whether they match. A user pronunciation determination device, comprising: determination means for determining a user pronunciation. 2. The user pronunciation according to claim 1, wherein said determination means is configured to teach a difference between said recognized character string and said set character string. Judgment device. 3. The user pronunciation determination according to claim 1, wherein the determination unit is configured to teach a degree of coincidence between the recognized character string and the set character string. apparatus. 4. A repetition means for causing a user to repeatedly execute a process of uttering the set character string until the recognized character string matches the set character string. Item 4. The user pronunciation determination device according to any one of Items 1 to 3. 5. A recording medium for recording a program for operating a user pronunciation determination device, the program comprising: a function as character string setting means for setting a character string for uttering a user; A function as voice recognition means for recognizing a character string corresponding to the voice by reading the character string and uttering the voice and performing voice recognition processing; and a display means for displaying the character string recognized by the voice recognition means. And a function as a determination unit for comparing the recognized character string with the set character string to determine whether or not they match each other, Recording medium.