JP2008046570A - Voice input system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform an instruction simply and accurately by a voice input. <P>SOLUTION: A voice recognition processing part 104 performs voice recognition of a voice data input by a voice input part 102 in a normal mode and a confirmation mode. An instruction accepting part 106 presumes the content of the instruction from a user based on a result of the voice recognition in normal mode performed by the voice recognition processing part 104 and shifts the voice recognition processing part 104 to the confirmation mode. When voice data are input from the voice input part 102 while the mode is shifted to the confirmation mode, the voice recognition processing part 104 performs voice recognition of the voice data in the confirmation mode. The instruction accepting part 106 detects whether or not the user has confirmed that the presumption was correct based on a result of the voice recognition in the confirmation mode by the voice recognition processing part 104, and determines the content of the instruction to accept the content as the instruction when it is confirmed that the presumption is correct. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a voice input system.

Conventionally, a control system capable of inputting operation commands by voice has been developed. Japanese Unexamined Patent Application Publication No. 2006-33795 describes a remote control system using voice input. In the technique described in this document, when there is a voice input from a remote control terminal, a voice recognition unit recognizes the voice and extracts a keyword. Based on this keyword, control items that can be selection candidates are searched. This control item is displayed on the television screen. The user can select and instruct a desired control item from among the displayed control items by using the remote control terminal as a pointing device. When the user operates the selection key of the operation key, the control code of the control item is acquired and transmitted to the television. Thereby, the target operation command can be input by a simple operation.
JP 2006-33795 A

  However, in the technique described in Patent Document 1, it is necessary to operate an operation key or the like of a remote control terminal in order to select an operation command after performing voice input. Therefore, for example, when the user has a physical disability and can only perform voice input, there is a problem that the operation cannot be performed.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique in which a user gives instructions simply and accurately by voice input.

According to the present invention,
A voice input unit for inputting voice data of the user's voice;
A speech recognition processing unit that recognizes speech data input by the speech input unit in at least two modes of a normal mode and a confirmation mode;
An instruction receiving unit that estimates the instruction content from the user and shifts the voice recognition processing unit to the confirmation mode based on a result of the voice recognition processing unit performing voice recognition in the normal mode;
In order to present the estimation by the instruction receiving unit and to cause the user to utter the correctness of the estimation, a presentation processing unit that presents the content to be uttered by the user;
Including
When voice data is input from the voice input unit during the transition to the confirmation mode, the voice recognition processing unit performs voice recognition by comparing the voice data with the content presented by the presentation processing unit,
The instruction receiving unit detects whether or not the user has confirmed that the estimation is correct based on a result of the voice recognition processing unit performing voice recognition in the confirmation mode, and it has been confirmed that the estimation is correct. In some cases, a voice input system is provided that accepts the instruction content and accepts it as an instruction.

  According to this configuration, the instruction is estimated based on the voice recognition result of the user's voice, the estimation result is presented, and the user is again input by voice whether or not it is correct. Since the second input increases the recognition accuracy, an instruction from the user can be accurately received only by voice. According to the voice input system of the present invention, since an instruction can be given only by voice input, the user can instruct a desired process only by speaking the necessary information.

  The voice input system of the present invention may further include a storage unit that stores the content presented by the presentation processing unit, and the voice recognition processing unit refers to the storage unit in the confirmation mode, and The voice data input from the voice input unit can be compared with the content.

  In the voice input system of the present invention, the instruction receiving unit estimates a content of the instruction from the user based on a result of the voice recognition processing unit performing voice recognition in the normal mode, and indicates a result of the estimation. A term can be determined, and the presentation processing unit can present the general term determined by the instruction receiving unit as the content that causes the user to speak.

  It should be noted that any combination of the above-described constituent elements and a conversion of the expression of the present invention between a method, an apparatus, a system, a recording medium, a computer program, etc. are also effective as an aspect of the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, a user can perform an instruction | indication simply and accurately by voice input.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same reference numerals are given to the same components, and the description will be omitted as appropriate.

(First embodiment)
FIG. 1 is a block diagram showing a configuration of a voice input system 100 according to the present embodiment.
The voice input system 100 includes a voice input unit 102, a voice recognition processing unit 104, an instruction receiving unit 106, a presentation processing unit 112, a processing unit 114, a mode storage unit 116, a presentation content storage unit 118, a voice data storage unit 120, and a processing content. A storage unit 122, a microphone 130, a speaker 132, and a display 134 are included. The voice input system 100 can be configured by, for example, a personal computer (hereinafter referred to as a PC). The microphone 130, the speaker 132, and the display 134 can be configured to be built in the PC or connected by wire or wirelessly.

  In the present embodiment, voice input system 100 receives a voice instruction from a user. The processing unit 114 performs various processes in accordance with instructions received from the user by the instruction receiving unit 106. Although the process of the process part 114 is not specifically limited, For example, insurance contract support processes, such as car insurance, can be performed. The processing content storage unit 122 stores candidates for processing performed by the processing unit 114. A specific configuration of the processing content storage unit 122 will be described later.

  The microphone 130 is provided in a place where the user's voice can be acquired. The voice input unit 102 inputs an instruction from the user by voice through the microphone 130. In the present embodiment, the voice input unit 102 digitally converts the voice input from the microphone 130. The audio data storage unit 120 stores audio data (hereinafter simply referred to as audio data) that has been input and digitally converted by the audio input unit 102.

  The voice recognition processing unit 104 recognizes the voice data input by the voice input unit 102 in at least two modes: a normal mode and a confirmation mode. Although not shown, the speech recognition processing unit 104 has a dictionary or the like included in a general speech recognition program. The normal mode is a voice recognition process similar to a general voice recognition program. The confirmation mode will be described later.

  The mode storage unit 116 stores the mode setting of the voice recognition processing unit 104. In the initial state, the mode storage unit 116 is set to the normal mode.

  The instruction receiving unit 106 estimates the instruction content from the user based on the result of the voice recognition processing unit 104 performing voice recognition in the normal mode. The processing content storage unit 122 stores instruction content candidates instructed by the user. The instruction receiving unit 106 refers to the processing content storage unit 122 based on the result of the voice recognition processing unit 104 performing voice recognition in the normal mode, and estimates the instruction content from the user from the instruction content candidates.

  The instruction receiving unit 106 estimates the instruction content from the user and shifts the voice recognition processing unit 104 to the confirmation mode. Specifically, the instruction receiving unit 106 changes the setting of the mode storage unit 116 from the normal mode to the confirmation mode, thereby causing the voice recognition processing unit 104 to shift to the confirmation mode.

  The presentation processing unit 112 performs a process of presenting the content to be uttered by the user in order to present the estimation by the instruction receiving unit 106 and to cause the user to utter the correctness of the estimation. The presentation processing unit 112 refers to the presentation content storage unit 118 and performs a process of presenting them to the user through the speaker 132 or the display 134.

  FIG. 2 is a diagram illustrating an example of an internal configuration of the processing content storage unit 122. The processing content storage unit 122 includes an instruction content field and a presentation content field. The presentation content column includes a confirmation content column, a positive column, and a negative column.

  As shown in FIG. 2A, for example, “contract processing start” and “re-input” are stored in the instruction content column. Corresponding to the instruction content “Contract processing start”, “Contract processing starts” as confirmation content, “Yes” as content to be uttered by the user when affirming the estimation, and denying the estimation “No” is stored as the content to be uttered by the user. The processing content storage unit 122 can further store a plurality of instruction contents.

  Returning to FIG. 1, the instruction receiving unit 106 stores the content to be uttered by the user and the estimated instruction content in the presentation content storage unit 118 in order to make the user utter the correctness of the estimated instruction content. That is, the same content as the presentation content column shown in FIG. 2A is stored in the presentation content storage unit 118.

  When voice data is input from the voice input unit 102, the voice recognition processing unit 104 refers to the mode storage unit 116 and confirms the currently set mode. When voice data is input from the voice input unit 102 during the transition to the confirmation mode, the voice recognition processing unit 104 compares the voice data with the content presented by the presentation processing unit 112 and performs voice recognition. Specifically, the voice recognition processing unit 104 refers to the presentation content storage unit 118 and compares the voice data input from the voice input unit 102 with the content stored in the presentation content storage unit 118. In the present embodiment, in the confirmation mode, since the content of the user's utterance can be specified to some extent, voice recognition can be performed with high accuracy.

  The instruction receiving unit 106 detects whether or not the user has confirmed that the estimation is correct based on the result of the voice recognition processing unit 104 performing voice recognition in the confirmation mode, and instructs when the estimation is confirmed to be correct. Confirm the content and accept it as an instruction. When the instruction receiving unit 106 determines the instruction content, the processing unit 114 performs processing based on the instruction.

  FIG. 3 is a flowchart showing a processing procedure of the voice input system 100 according to the present embodiment.

  When the voice input unit 102 inputs voice data (YES in S100), the voice recognition processing unit 104 refers to the mode storage unit 116 and checks whether the mode is the normal mode or the confirmation mode (S102). When the normal mode is set (YES in S102), the voice recognition processing unit 104 performs voice recognition processing in the normal mode (S104).

  The instruction receiving unit 106 refers to the processing content storage unit 122 based on the voice recognition result of the voice recognition processing unit 104 and estimates the user's instruction content (S106). The instruction receiving unit 106 stores the estimated instruction content and a term indicating whether or not to affirm the estimation in the presentation content storage unit 118 (S108). In addition, the instruction receiving unit 106 rewrites the setting in the mode storage unit 116 from the normal mode to the confirmation mode. Thereby, the speech recognition processing of the speech recognition processing unit 104 is shifted to the confirmation mode (S110).

  The presentation processing unit 112 refers to the presentation content storage unit 118 and performs a process of presenting the instruction content estimated by the instruction receiving unit 106 and a term indicating whether or not to affirm the estimation using the speaker 132 or the display 134 ( S112).

  Thereafter, when the voice input unit 102 receives a voice input (YES in S114), the process proceeds to step S102. In this case, in step S102, the confirmation mode is determined (NO in S102), and the speech recognition processing unit 104 performs speech recognition processing in the confirmation mode (S120). The instruction receiving unit 106 determines whether or not the instruction estimated in step S106 has been affirmed by the user based on the voice recognition result of the voice recognition processing unit 104 (S122). When the estimation is affirmed (YES in S122), the instruction receiving unit 106 confirms the estimated instruction and receives the instruction (S124). The processing unit 114 starts processing based on the instruction determined by the instruction receiving unit 106 (S125).

  On the other hand, if the estimation is not affirmed in step S122 (NO in S122), the instruction receiving unit 106 determines whether or not the estimation is denied (S126). If the estimation is denied (YES in S126), a cancel process is performed (S130). The canceling process means returning to the state before the voice input process of step S100 is performed. Further, the instruction receiving unit 106 rewrites the setting in the mode storage unit 116 from the confirmation mode to the normal mode. At this time, the instruction receiving unit 106 may further perform processing for deleting the estimation of the instruction content stored in the presentation content storage unit 118.

  If the estimation is not denied in step S126 (NO in S126), that is, if the voice input in step S114 is not correctly recognized, the instruction receiving unit 106 checks the setting of the mode storage unit 116 in the confirmation mode. To normal mode (S128). At this time, the instruction receiving unit 106 may further perform processing for deleting the estimation of the instruction content stored in the presentation content storage unit 118. Thereafter, the process returns to step S102, and a process of recognizing the voice input in step S114 in the normal mode is performed (S104). The voice recognition processing unit 104 performs voice recognition of the voice data stored in the voice data storage unit 120.

  On the other hand, after step S112, if there is no voice input even after a predetermined time has elapsed (NO in S114, YES in S116), the same canceling process as in step S130 is performed (S118). The process ends as described above.

  Next, a specific example will be described. Here, it is assumed that the mode storage unit 116 is set to the normal mode. For example, it is assumed that when a user performs some contract procedure, the user speaks “I will contract”. The voice recognition processing unit 104 performs voice recognition processing of voice data “I will contract” in the normal mode.

  The instruction receiving unit 106 refers to the processing content storage unit 122 based on the voice recognition result by the voice recognition processing unit 104 and estimates that the instruction content from the user is “contract processing start”. The instruction receiving unit 106 refers to the processing content storage unit 122 and reads the content to be presented by the presentation processing unit 112. At this time, when the setting is made as shown in FIG. 2 (a), “contract processing start” as the instruction content, “contract processing starts” as the confirmation content, The presentation content storage unit 118 stores “Yes” as the content to be uttered and “No” as the content to be uttered by the user when the estimation is denied. In addition, the instruction receiving unit 106 rewrites the setting in the mode storage unit 116 from the normal mode to the confirmation mode.

  The presentation processing unit 112 refers to the presentation content storage unit 118 and says “Start contract processing. If yes, say“ Yes ”, otherwise say“ No ”. "Is presented to the user through the speaker 132 or the display 134.

  In this state, for example, when the user utters “Yes”, the voice recognition processing unit 104 performs voice recognition processing in the confirmation mode. At this time, since the voice input from the user is likely to be “Yes” or “No”, the voice recognition processing unit 104 can perform voice recognition with high accuracy. The instruction receiving unit 106 detects that the instruction “contract processing start” has been confirmed by the user based on the voice recognition result by the voice recognition processing unit 104, and receives the instruction “contract processing start”. On the other hand, when it is not possible to detect that the user has spoken “Yes”, or when it has been detected that the user has spoken “No”, cancel processing is performed. Furthermore, when the user's speech content cannot be detected, the speech recognition process is performed again in the normal mode. Thereafter, the same processing is repeated according to the voice recognition result.

  The instruction receiving unit 106 detects whether or not the user has uttered that the result is correct with respect to the instruction presented by the presentation processing unit 112, and when it is detected that the user has uttered that the result is correct, Confirm the instruction. Thereafter, contract processing is started.

  Next, another specific example will be described. The instruction receiving unit 106 can estimate the instruction content from the user based on the result of the voice recognition processing unit 104 performing voice recognition in the normal mode, and can determine a general term indicating the estimation result. The presentation processing unit 112 can present general terms determined by the instruction receiving unit 106 as content that causes the user to speak. An example of the internal configuration of the processing content storage unit 122 when performing operations in this example is shown in FIG. Again, for example, “contract processing start” is stored in the instruction content column. In addition, “Please start contract processing” is stored as a general term for the instruction content of contract processing start. That is, in association with this instruction content, “contract processing starts” is stored as confirmation content, “start contract processing” is stored as an affirmative, and “different” is stored as negative.

  First, it is assumed that the mode storage unit 116 is set to the normal mode. For example, it is assumed that when a user performs some contract procedure, the user speaks “I will contract”. The voice recognition processing unit 104 performs voice recognition processing of voice data “I will contract” in the normal mode.

  The instruction receiving unit 106 estimates that the user is instructing “contract processing start” based on the voice recognition result by the voice recognition processing unit 104. The instruction receiving unit 106 refers to the processing content storage unit 122 and reads the content to be presented by the presentation processing unit 112. At this time, if the setting is as shown in FIG. 2B, the instruction content is “Contract processing start”, the confirmation content is “Start contract processing.” The presentation content storage unit 118 stores “Please start contract processing” as the content to be uttered and “No.” as the content to be uttered by the user when the estimation is denied. In addition, the instruction receiving unit 106 rewrites the setting in the mode storage unit 116 from the normal mode to the confirmation mode.

  The presentation processing unit 112 refers to the presentation content storage unit 118 and says “Start contract processing. If you like, please start contract processing.” If not, please say “No.” . "Is presented to the user through the speaker 132 or the display 134.

  In this state, for example, when the user speaks “Please start contract processing”, the speech recognition processing unit 104 performs speech recognition processing in the confirmation mode. At this time, since the voice input from the user is likely to be “Please start contract processing” or “No.”, the voice recognition processing unit 104 can perform voice recognition with high accuracy. . The instruction receiving unit 106 detects that the instruction “contract processing start” has been confirmed by the user based on the voice recognition result by the voice recognition processing unit 104, and receives the instruction “contract processing start”.

  On the other hand, when it is not detected that the user has uttered “Please start contract processing”, or when it is detected that the user has uttered “No”, cancel processing is performed. Furthermore, when the user's speech content cannot be detected, the speech recognition process is performed again in the normal mode. Thereafter, the same processing is repeated according to the voice recognition result.

  According to voice input system 100 in the present embodiment, an instruction can be given only by voice input, and thus the user can instruct a desired process only by speaking necessary information.

(Second Embodiment)
This embodiment is different from the first embodiment in that a microphone is provided in the remote control terminal.

FIG. 4 is a diagram showing a configuration of the voice input system 100 according to the present embodiment.
In the present embodiment, voice input system 100 includes a system main body 101 and a remote control terminal 200. The system main body 101 has the same configuration as that of the voice input system 100 according to the first embodiment except that the system main body 101 does not have the microphone 130 but has the receiving unit 150.

  The remote control terminal 200 includes a microphone 202, an audio conversion unit 204, and a transmission unit 206. The voice conversion unit 204 digitally converts the voice input from the microphone 202. The transmission unit 206 transmits the audio data converted by the audio conversion unit 204 to the reception unit 150 of the system main body 101. The voice input unit 102 inputs the voice data received by the receiving unit 150. Communication between the reception unit 150 and the transmission unit 206 can be performed using various networks, but can be performed using, for example, infrared rays.

  4 shows a configuration in which the system main body 101 includes the speaker 132 and the display 134, these may be provided separately from the system main body 101. Further, either one or both of the speaker 132 and the display 134 may be provided in the remote control terminal 200. Further, other than the speaker 132 and the display 134, each component of the system main body 101 does not need to be provided in one casing, and is provided in a plurality of terminals so as to transmit / receive data via a network or the like. It is good also as a structure to perform.

  Each component surrounded by a broken line in the voice input system 100 shown in FIGS. 1 and 4 is not a hardware unit configuration but a functional unit block. Each component surrounded by a broken line of the voice input system 100 includes an arbitrary computer CPU, memory, a program for realizing the components shown in the figure loaded in the memory, a storage unit such as a hard disk for storing the program, and a network connection. It is realized by any combination of hardware and software, with a focus on the interface. It will be understood by those skilled in the art that there are various modifications to the implementation method and apparatus.

  As mentioned above, although embodiment of this invention was described with reference to drawings, these are the illustrations of this invention, Various structures other than the above are also employable.

  In the above embodiment, the configuration in which the voice input system 100 includes the presentation content storage unit 118 and the processing content storage unit 122 has been described, but these may be integrally formed. In this case, the instruction receiving unit 106 can distinguish the instruction contents estimated from the user from among the plurality of instruction contents stored in the processing content storage unit 122 by marking or the like.

  In the above embodiment, the voice input system 100 includes the speaker 132 and the display 134. However, the voice input system 100 may include only one of the speaker 132 and the display 134.

It is a block diagram which shows the structure of the audio | voice input system in embodiment of this invention. It is a figure which shows an example of an internal structure of a process content storage part. It is a flowchart which shows the process sequence of the audio | voice input system in embodiment of this invention. It is a block diagram which shows the structure of the audio | voice input system in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Voice input system 101 System main body 102 Voice input part 104 Voice recognition process part 106 Instruction reception part 112 Presentation process part 114 Process part 116 Mode storage part 118 Presentation content storage part 120 Voice data storage part 122 Process content storage part 130 Microphone 132 Speaker 134 Display 150 Reception Unit 200 Remote Control Terminal 202 Microphone 204 Audio Conversion Unit 206 Transmission Unit

Claims (3)

A voice input unit for inputting voice data of the user's voice;
A speech recognition processing unit that recognizes speech data input by the speech input unit in at least two modes of a normal mode and a confirmation mode;
An instruction receiving unit that estimates the instruction content from the user and shifts the voice recognition processing unit to the confirmation mode based on a result of the voice recognition processing unit performing voice recognition in the normal mode;
In order to present the estimation by the instruction receiving unit and to cause the user to utter the correctness of the estimation, a presentation processing unit that presents the content to be uttered by the user;
Including
When voice data is input from the voice input unit during the transition to the confirmation mode, the voice recognition processing unit performs voice recognition by comparing the voice data with the content presented by the presentation processing unit,
The instruction receiving unit detects whether or not the user has confirmed that the estimation is correct based on a result of the voice recognition processing unit performing voice recognition in the confirmation mode, and it has been confirmed that the estimation is correct. A voice input system that accepts the instruction content as an instruction in a case. The voice input system according to claim 1,
A storage unit for storing the content presented by the presentation processing unit;
In the confirmation mode, the voice recognition processing unit refers to the storage unit and compares the voice data input from the voice input unit with the content. The voice input system according to claim 1 or 2,
The instruction receiving unit estimates the instruction content from the user based on a result of the voice recognition processing unit performing voice recognition in the normal mode, determines a general term indicating a result of the estimation,
The presentation processing unit is a voice input system that presents the general term determined by the instruction receiving unit as the content that causes the user to speak.

Images