JP2002251199A - Voice input information processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To actualize a voice input information processor which prevents input ted voice information itself or its voiceprint from leaking out. SOLUTION: This voice input information processor has a voice input means 11gh equipped with a voice detecting means 11g for detecting an input voice, a sound signal generating means which generates a sound signal actively muting a voice to the outside by generating an opposite-phase sound from the detected voice or a sound signal for actively changing the voiceprint by varying the frequency spectrum of the input voice on the basis of a voice mode set in advance, and a sound output means 11ij which outputs the sound signal; and a signal regarding the detected input voice is transmitted to a CPU 11a to carry out an indicated process.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice input information processing device for performing authentication and operation instructions by a voice input device of an information processing device.

[0002]

2. Description of the Related Art At present, there has been proposed a system using voice input to instruct an information processing apparatus (computer) to perform various processes. There is also an information processing apparatus that uses individual differences in the characteristics of human voice prints as authentication information for authentication. JP-A-9-146
No. 563 is an example of using this voiceprint. However, it is possible for an external person to record a human voiceprint input at the time of authentication with another voice recording device and steal the password information.

[0003] Active silencers for reducing voice or noise have also been proposed. However, a method combining these has not been proposed. However, voice authentication is easy to steal the voice and requires some protection. In contrast, Japanese Patent Application Laid-Open No. 08-08
Japanese Patent No. 4190 proposes a method of preventing eavesdropping leakage by simple recording by superimposing a dummy push-phone signal on a push-dial line at the time of authentication in the case of telephone authentication. The signal is superimposed and can be easily removed. Also, it can be applied only to telephone lines.

[0004] A system such as Japanese Unexamined Patent Application Publication No. 10-092121 has also been proposed in which sound (noise) other than the input sound is reduced by using active muffling means. Does not disclose an example using an active silencer. Further, as disclosed in Japanese Patent Application Laid-Open No. 7-170587, there is disclosed a voice input device which detects the utterance of a sounder with a microphone and actively mutes the sound based on the detected signal. This is an active silencer that does not relate to authentication and processing instructions, but simply cancels sound by actively canceling voice + sound input to a microphone. Conventionally, the authentication and the processing instruction by the voice input of the information processing apparatus leak the voice to the outside.

[0005]

At present, a system using voice input has been proposed to instruct an information processing apparatus to perform various processes. There is also an information processing apparatus that uses individual differences in the characteristics of human voice prints as authentication information for authentication. However, it is possible for an external person to steal personal identification information by recording the human voiceprint input at the time of authentication by another voice recording device. The present invention has been made in view of such circumstances, and a first object of a voice input information processing apparatus according to the present invention is to provide a voice input unit having a function of actively silencing a voice so that a voice can be externally heard. An object of the present invention is to prevent leakage of personal identification information at the time of authentication by voice by means of a method of eliminating voice information or voice input means having a function of actively changing a voiceprint. A second object of the present invention is to prevent a malfunction caused by a voice of an adjacent operator and quiet the office when a large number of operators perform operations by voice input as in a normal office. That is, the objects in the claims according to the present invention are as follows.

[0006] The object of the present invention is to actively cancel the voice of the operator to prevent leakage of the voice to the outside. Also,
Make the office quieter by not hearing any sounds around.

[0007] The voiceprint of the operator's voice is actively changed to prevent the voiceprint from leaking to the outside even if the voice information can be heard outside.

A third object of the present invention is to realize voice authentication which can perform authentication without leaking a voice or a voiceprint to the outside.

[0009] (Object of Claim 4) Since the operation instruction of the information processing apparatus is performed by a voiceprint having the same characteristics as the authenticated voice, even if there are a plurality of sounders, the voice operated by the instruction of only a specific operator is provided. An information processing device having input means is realized.

[0010] (Object of claim 5) Since the operation instruction of the information processing apparatus is performed by using a voiceprint having the same characteristics as the authenticated voice, even if there are a plurality of sounders, the voice operated by the instruction of only a specific operator is provided. An information processing apparatus including an input unit is realized, and leakage of personal identification information at the time of authentication is prevented.

(Purpose of Claim 6) The information processing device displays or sounds pronunciation information for authentication, and authenticates the password information for the authentication by authenticating based on the voiceprint information of the voice reproduced therefrom. Need not be remembered,
Authentication is assured by voiceprint without leaking to the outside.

[0012]

According to a first aspect of the present invention, there is provided an audio processing apparatus comprising: a voice detecting means for detecting a voice; and a sound generating means for generating a sound having an opposite phase from the detected voice. The voice input means has a voice input means having an active muffling function to mute the sound, and the voice input means transmits a signal related to the detected voice to the information processing device and prevents the input voice from leaking outside. The voice input information processing device performs the following.

According to a second aspect of the present invention, there is provided a voiceprint changing function of actively changing a voiceprint by voice detecting means for detecting voice and sound generating means for generating a sound based on the detected voice. By providing a voice input unit, the voice input unit transmits a signal related to the detected voice to an information processing device, and also allows a voice of a different voiceprint to be heard outside. It is characterized by the following.

According to a third aspect of the present invention, in the voice input information processing apparatus according to the first or second aspect, the voice input means performs authentication by extracting a voiceprint feature of the voice. It is characterized by being a device.

According to a fourth aspect of the present invention, in the voice input information processing apparatus provided with voice detection means for detecting voice and voiceprint detection means for detecting voiceprint based on the detected voice, the authentication is performed. A voice input information processing apparatus that recognizes only an operation command by voice of a voiceprint having the same characteristics as the voiceprint and receives an operation.

According to a fifth aspect of the present invention, in the voice input information processing apparatus according to the third aspect, a voice detecting means for detecting a voice, and a voiceprint detecting means for detecting a voiceprint based on the detected voice. Wherein the voice input information processing apparatus recognizes only an operation command by voice of a voice print having the same characteristics as the authenticated voice print, and receives the operation. Things.

According to a sixth aspect of the present invention, in the voice input information processing apparatus according to the third aspect, at the time of authentication, the voice input information processing apparatus displays or sounds pronunciation information for authentication, and A speech input information processing device that pronounces pronunciation information and performs authentication based on a voiceprint of the pronounced speech.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention broadly relates to a voice input means having two types of an active mute function and an active change function, and a means for authenticating the voice input means and giving a processing operation instruction of an information processing apparatus. It is. Appropriate authentication or operation instruction is realized by adopting a configuration in which a voice input unit having two types of active silence function or active change function and authentication or operation instruction unit in the information processing apparatus are used in combination. I do.

First, two types of voice input means will be described. Both of them actively change the voice of the speaker and can be realized by the configuration as shown in FIG. here,
FIG. 1 is a conceptual diagram showing one embodiment of a voice input unit in a voice input information processing device according to the present invention. First, the sound of the sounder 10 is detected by the sound detecting means 1, and the detected sound signal is input to the sound signal creating means 3 and the sound transmitting means 4. The sound signal creating means 3 creates an interference sound signal for canceling the input audio signal or changing the audio signal, and outputs the created interference sound signal to the sound generating means 2.
And the sound generating means 2 generates an interference sound. Generally, sound including sound is a wave due to the vibration of air, and as shown in FIG. 3, when a sound wave having an opposite phase to the sound is caused to interfere, the interfering synthetic sound wave (synthetic sound) is canceled by the interfering sound. . Basically, the active silencer is configured based on this principle.

The configuration example shown in FIG. 1 is a case where a pair of voice detection means 1 and sound generation means 2 are provided. As shown in FIG. 2, a large number of pairs of voice detection means 1, 1 ', 1 "and sound generating means 2, 2 ', 2" may be used. In this example, the sound detection means 1, 1 ', 1 "give the sounder 10 a directivity, while the sound generation means 2,
2 ′, 2 ″ are provided with directivity on the outside on the side opposite to the sounder 10 so that many sound generating means 2, 2 ′,
On the inside of 2 ", that is, on the side of the sounder 10, the sound of the sounder 10 can be heard, but on the outside of the many sound generating means 2, 2 ', 2", that is, on the opposite side of the sounder 10, the sound is canceled.
Leakage of the sound of the sounder 10 can be prevented.

In the above description, the method of canceling the voice has been described. Next, the method of changing the voiceprint will be described. Generally, a voiceprint of a human voice recognizes an individual by focusing on the fact that the distribution of the frequency spectrum of a generated sound differs between individuals, and on the strength and interval of vowels and consonants. The distribution of the frequency spectrum of the generated sound is caused by individual differences such as the vocal cords of the human, the size of the oral cavity, and the opening and closing operations of the mouth. Therefore, in particular, there are individual differences, and vocal cords that are difficult for humans to consciously change, by canceling or changing the frequency spectrum portion caused by the size of the oral cavity,
The voiceprint can be changed while leaving the information as voice. That is, for example, when the frequency spectrum of the original voice of the sounder is as shown in FIG. 4A, the frequency spectrum of a part of the original voice is changed so that the frequency of the changed synthesized voice after the interference is changed. The spectrum is as shown in FIG.
, The voiceprint can be changed.

In the example shown in FIG. 4B, an example is shown in which some frequency spectrum bands are canceled. However, a plurality of frequency spectrum bands may be changed, or the frequency spectrum band may be changed over time. You may do it. In this case, if the frequency spectrum band to be changed is dispersed,
Since the voice change due to the voiceprint change becomes difficult for the human ear to understand, there is little discomfort with using such a voiceprint change device, and since the frequency to be changed extends over a plurality of frequency spectrum bands, It is possible to make it difficult to analyze a voiceprint from a synthesized voice that can be heard outside.

Next, FIG. 5 shows a basic configuration in which the active mute function or the active voiceprint changing function as described above is applied as a voice input means of a voice input information processing apparatus. Here, voice input means 5 with an active mute (change) function
Has the sound detecting means 1, the sound generating means 2, the sound signal creating means 3, and the sound transmitting means 4 shown in FIG. 1, and the sound generating means 2 has an active silencing function as described above. Alternatively, it has an active voiceprint changing function.

First, voice input means 5 having an active silence function
In the case of, the sound produced by the sounder does not leak to the outside, and the speech input means 5 transmits the created speech signal to the information processing device 6. The information processing device 6 performs various operations based on the transmitted audio signal. On the other hand, in the case of the voice input means 5 with the active voiceprint changing function, the voice of the voiceprint changed from the original voice pronounced by the speaker leaks to the outside, but the original voice pronounced by the speaker does not leak outside. The voice input means 5 transmits the generated voice signal to the information processing device 6. The information processing device 6 performs various operations based on the transmitted audio signal.

In order to ensure security, the current information processing device clarifies the user of the information processing device and limits the range in which the user can issue an instruction to the information processing device. In general, as shown in the authentication operation flowchart of FIG. 6, an authentication operation for clarifying a user's identity to the information processing device before the user performs various operations on the information processing device. Only after the authentication is obtained, the user can perform various operations within the authority of the user.

That is, as shown in the flowchart of FIG. 6, usually, in the authentication operation, the information processing apparatus requests the user to input a user name and a password, and in response to this, the user You are authenticated by entering your name and PIN. After being authenticated, the user can perform various operations within the authority of the user. The user,
When the operation on the information processing apparatus is no longer necessary, the end of the operation is input to the information processing apparatus, and the operation is terminated. Can be used.

However, conventionally, in the authentication operation as shown in FIG. 6, there are the following two problems. The first problem is
If the user of the information processing apparatus needs to constantly memorize the user name and the password, and if another user illegally obtains the user name and the password by some method, In other words, the information processing apparatus can be used by impersonating the user. The second problem is
After the authentication operation, if the user is operating the information processing apparatus, or leaves the information processing apparatus without inputting the end of the operation to the information processing apparatus, another user may operate the information processing apparatus. Also, since the information processing device is indistinguishable from an authenticated user or another user, even other users,
The information processing device can be used. The configuration of the voice input information processing apparatus according to the present invention can solve such a problem.

First, FIG. 7 shows a specific configuration example of a voice input information processing apparatus provided with an active mute (change) function voice input means for realizing the function of the present invention. In FIG. 7, a central processing unit (CPU) 11a includes a voice input information processing device 1
1 and performs various processes. The main storage (RAM) 11b is a part for storing data and programs for performing various processes. Auxiliary storage (HD)
11c is also a part for storing data and programs, but the main storage unit 11b stores information when the power of the voice input information processing device 11 is turned off, while the stored information is lost. The unit 11c uses a hard disk device or the like, and the power of the voice input information processing device 11 is turned off.
Even if it is set to F, it is possible to hold the stored information. The display unit (CRT) 11d is a means for transmitting information to the user by displaying the information from the voice input information processing device 11 to the user. The keyboard 11e is a means by which the user transmits information to the voice input information processing device 11. The communication unit 11f transmits information to and from another information processing device using a cable, wireless communication, or the like.

The above is an example of the configuration of a general information processing apparatus. In the voice input information processing apparatus 11 according to the present invention, a microphone 11g for converting a voice into a voice signal.
And an A / D converter 11h with an amplifier for amplifying and then A / D converting the audio signal to a digital signal, and an amplifier for converting digital audio information from the CPU into an analog audio signal. D / A converter 11i
And a speaker 11j for converting the analog sound signal into sound. Here, the sound input means 11gh is for inputting sound, and the sound output means 11ij is for actively silencing or changing the voiceprint to the outside, as described above.

Voice input means 11gh, sound output means 11
ij is not limited to a pair, and as shown in FIG. 2, a plurality of voice input means 11gh, 11'gh and sound output means 11ij, 1 are provided so that the voice of the sounder does not leak outside.
1′ij may be arranged efficiently. The sound output means 11ij may be used for the purpose of transmitting information from the information processing device to the user by sound, in addition to actively silencing the sound or changing the voice print.

Next, the flow of the entire processing in the voice input information processing apparatus according to the present invention is shown in the flowchart of FIG. First, when the power of the voice input information processing apparatus 11 is turned on, after checking the operation of various circuit blocks, the voice input information processing apparatus 11 loads a startup program from the auxiliary storage unit 11c into the main storage unit 11b, and performs voice input processing. After activating the operation system OS, the information processing apparatus 11 enters a “start” state indicating a wait for use by a user. Thereafter, as described above, various processing routines such as a mute operation, a voiceprint change operation, a voiceprint determination operation, and a voice recognition operation according to the present invention operate concurrently with the main routine shown in FIG. .

FIG. 12 shows an example in which a timer interrupt is used as an example for realizing such simultaneous and parallel operations. After activation, the voice input information processing device 11
Even when the main routine S100 is executed, the timer interrupt signal 100a is generated at a fixed time interval 100b, and after the processing of the main routine S100 is interrupted, the process jumps to the timer interrupt routine S200. After the end of the interruption routine S200, the process returns to the original interruption point of the main routine S100, and the processing of the main routine S100 is continued from the interruption point. By adopting such a method, other processing can be executed in parallel with the main routine S100.

Therefore, by incorporating a processing routine for performing various operations as shown in FIG. 11 into an interrupt routine by a timer, the above-described mute (change) operation, voiceprint detection, or voiceprint detection can be performed in parallel with the processing of the main routine. The operation of voice recognition can be executed. Here, FIG.
9 is a flowchart showing the flow of an interrupt routine for performing a mute (change) operation, a voiceprint detection operation, and a voice recognition operation. First, the input voice information is obtained from the voice input unit 11gh (step S201).

Next, it is determined whether the preset sound signal creation mode is the mute mode or the voiceprint change mode is specified (step S202), and the mute mode or the voiceprint change mode is set. If there is (YES in step S202), the designated sound signal is created (step S203). The acoustic signal to be created is, as described above, an interference signal that generates a muffling signal that cancels the input voice or changes the frequency spectrum. The interference signal is output via the sound output unit 11ij (step S20).
4) Thus, in the case of the mute mode, the input voice can be canceled so as not to leak outside. In the case of the voiceprint change mode, as described above, an interference signal that cancels a part of the frequency spectrum is output,
It is possible to prevent the voiceprint of the input original voice from leaking outside.

Next, it is checked whether the voiceprint determination mode is set (step S205). If the voiceprint determination mode is set (YES in step S205), a voiceprint is detected (step S206). As for voiceprint detection, as disclosed in Japanese Patent Application Laid-Open No. Hei 5-210648, the feature amount of each individual voiceprint is recorded and compared with the feature amount extracted from the voice information input by pattern matching. Just do it.

Next, it is checked whether or not the voice recognition mode is set (step S207). If the voice recognition mode is set (YES in step S207), the operation is started.
As disclosed in Japanese Patent No. 3687, the voice may be recognized (step S208).

After the above processing is completed and the processing of the interrupt routine is completed, the processing of the main routine is resumed from the interrupted point and continued, whereby the mute operation, voiceprint change operation, voiceprint recognition operation performed in the interrupt routine are performed. ,
The voice recognition operation is executed in parallel with the processing of the main routine. It should be noted that as to the voiceprint recognition and voice recognition methods, more methods have been disclosed than ever, and an appropriate method may be used among them.

The processing of each of the above operations is performed by one CPU.
Although an example in which the processing is executed only by the CPU 11a has been described, in order to reduce the load on the CPU 11a, as shown in FIG. 13, a dedicated signal processing means (DSP) 12 is separately provided, which is realized as the interrupt routine. Various processes may be performed by the signal processing unit (DSP) 12. That is, in this case, the DSP 12 may be configured to transfer the result of performing each processing of the interrupt routine shown in FIG. 11 from the DSP 12 to the CPU 11a based on a certain interrupt signal from the CPU 11a.

Next, the flow of the entire processing (that is, the main routine) in the voice input information processing apparatus according to the present invention will be described with reference to the flowchart of FIG. In authenticating the user, first, what kind of authentication is to be performed is determined (step S101). In the present embodiment, either a mute authentication mode in which the user's voice is not leaked to the outside or a voiceprint change mode in which the voiceprint is not leaked to the outside can be selected.

In the case of the mute authentication mode (step S10
1), a mute authentication mode setting operation is performed (step S102), and as shown in FIG. 9, a mute mode (step S102a), a voiceprint determination mode (step 102b), and a voice recognition mode (step S102).
Set each flag of c).

On the other hand, in the case of the voiceprint change mode (in the case of the change mode in step S101), the setting operation of the voiceprint change authentication mode is executed (step S103), and as shown in FIG. 10, the voiceprint change mode (step S103a). And the voice print determination mode (step S103b) and the voice recognition mode (step S103c).

Based on the flags set in the mute authentication mode setting operation shown in FIG. 9 or the voiceprint change authentication mode setting operation shown in FIG. 10, as shown in FIG. The processing device 11 performs an operation related to the mode for which the flag is set in the interrupt routine S200.

In the case of the setting operation of the mute authentication mode shown in FIG. 9, an interruption routine S2 executed at regular intervals
In step S203, an interference sound is created so as to cancel the input voice signal so as not to leak the input voice to the outside (step S203), and the voiceprint feature amount is extracted from the voice of the input signal by voiceprint detection. (Step S206), voice is recognized by voice recognition to detect a user name and the like (Step S208). Thus, it is determined whether or not authentication is possible with the detected voiceprint or user name, and an authentication flag is set (step S102d).

On the other hand, the operation of setting the voiceprint change authentication mode shown in FIG. 10 is almost the same, and only the interference signal needs to cancel out a specific frequency spectrum of the voice (step S203). Alternatively, a characteristic frequency spectrum portion of another voiceprint may be canceled. As described above, after performing authentication based on the authentication information obtained by the mute authentication operation or the voiceprint change authentication operation and setting the authentication flag (step S102d or 103d), returning to FIG. (S104) When the authentication is OK (OK in step S104), the user enables the voice input information processing apparatus and logs in.

After logging in, the voice mode is determined (step S105). If the mode is the mute mode (step S105)
In the case of the mute mode, the mute mode flag is set (step S106), while in the case of the voiceprint change mode (the change mode of step S105), the voiceprint change mode flag is set (step S107). If the mute mode flag is set, a mute operation is performed in an interrupt routine started by a periodic interrupt by a timer (steps S203 and S20).
4). If the voiceprint change mode flag is set, similarly, the voiceprint is changed so that the voiceprint does not leak outside (steps S203 and S204).

Next, the operation mode is determined (step S108). In the case of the voiceprint determination mode, if the voiceprint extracted in the interrupt routine is authenticated as a user (OK in steps S109 and S110), the voiceprint determination is performed. Is OK, the operation command of the authenticated user is recognized from the voice by the voice recognition operation (step S111), and the voice input information processing device performs processing according to the operation command (step S1).
12) After the processing is completed, the audio mode is determined again (step S
Return to 105). The setting of the voice mode and the operation mode can be arbitrarily switched by the voice input information processing apparatus by a voice operation command.

That is, as an example, in the standby state after the authentication, the mute mode and the voiceprint determination mode are set to the initial state, so that the user does not leak his / her own voice to the information processing apparatus. Can be operated and after authentication,
Even if another person tries to operate illegally, the voiceprint determination becomes NG, and the other person cannot operate. Thereby, confidentiality can be maintained, and the office can be quieted. Also, P which constitutes the voice input information processing device
If it becomes necessary to have a conversation with another person while using C, by setting the voice mode to the voiceprint change mode, the conversation can be made with another person without silencing. When the voice input information processing device currently in use is used by another person with the consent of the person concerned, the operation mode may be set to the non-judgment mode. These mode changes can be performed only by the authenticator, thereby protecting the security.

For input information from other input means such as a keyboard and a pointing device, when the operation mode is set to the voiceprint determination mode, the voiceprint of the authenticator is detected for a certain period or more. If you cannot, do not accept input. Thus, when the user leaves the seat, the input from the other input means can be locked. Further, whether or not the user is in the seat may be detected using a human sensor or the like.

As described above, the voice input information processing apparatus using the voice input means having the active noise reduction function improves the security at the time of authentication and operation. Further, when a user name or the like is used as a voiceprint detection used at the time of authentication, the security is protected by the above-mentioned voiceprint changing means at the time of input of the voice input information processing apparatus, but the user name is being conversed in another place. However, in some cases, the user may be instructed to pronounce pronunciation information, for example, alphanumeric characters at random, and perform authentication based on the voiceprint of the pronounced sound. Also, as shown in FIG. 14, a voiceprint database 13 in which voiceprint feature amounts for user names are recorded is prepared for each user, and voiceprint feature amounts for various character strings are recorded. You may also learn in advance. Further, in order to further increase the security of the authentication, the authentication may be performed in combination with another authentication such as a fingerprint.

As described above, since the voice input means has a mute function or a voiceprint changing function, the input voice or the voiceprint of the input voice does not leak to the outside, and the voice input information processing device has Security can be maintained. In addition, since it is possible to authenticate based on the voiceprint information of the voice repetition of the voice reproduced in response to the display or pronunciation from the voice input information processing device, it is necessary for the authenticator to store the password information for the authentication. Since the authentication is performed by using a voiceprint without leaking to the outside, the authentication can be surely performed.

Further, since the operation instruction of the voice input information processing device is received only by a voiceprint having the same characteristic as the authenticated voice, even if there are a plurality of sounders, the operation can be performed by the instruction of only a specific operator. Further, even when the user is operating the voice input information processing apparatus or leaves the seat without inputting the end of the operation to the voice input information processing apparatus, another user cannot perform the operation by impersonating. Further, the authentication information does not leak outside.

[0052]

According to the first aspect of the present invention, since the voice input means having the mute function is provided, the voice does not leak around the operator. Therefore, there is no leakage of confidential information outside, and the office can be quietened.

(Effect of Claim 2) Since the voiceprint changing means is provided, the voiceprint is not leaked to the outside, and the voice can be transmitted to another person and conversation can be performed.

(Effect of Claim 3) Since authentication can be performed without leaking voice or voiceprint to the outside, voice authentication can be performed without leaking password information.

(Operation and Effect of Claim 4) Since the operation instruction of the voice input information processing apparatus is performed by the voiceprint having the same characteristic as the authenticated voice, even if there are a plurality of sounders, the instruction is performed only by the specific operator. Operable.

(Operation and Effect of Claim 5) Since the operation instruction of the voice input information processing apparatus is performed by the voice print having the same characteristics as the voice that has been authenticated, even if there are a plurality of sounders, the instruction of only the specific operator is given. And the authentication information does not leak outside.

(Operation and Effect of Claim 6) The voice input information processing device displays or sounds pronunciation information for authentication.
Authentication is performed based on the voiceprint information of the voice that is read back, so there is no need for the authenticator to memorize the password information for authentication, and since the authentication is performed using the voiceprint without leaking to the outside, authentication is ensured. It can be performed.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing one embodiment of a voice input unit in a voice input information processing device according to the present invention.

FIG. 2 is a conceptual diagram showing another embodiment of a voice input unit provided with a large number of pairs of voice detection units and sound generation units.

FIG. 3 is a schematic diagram showing a relationship between a synthesized sound signal of an input sound signal and an interference sound signal of an opposite phase.

FIG. 4 is a schematic diagram showing an example of a frequency spectrum when a frequency spectrum of an original voice of a sounder is changed.

FIG. 5 is a configuration diagram showing a basic configuration of a voice input information processing apparatus having voice input means to which an active mute function or an active voiceprint changing function is applied.

FIG. 6 is a flowchart illustrating a concept of an authentication operation of the information processing apparatus.

FIG. 7 shows a specific configuration example of an information processing apparatus provided with an active mute (change) function voice input unit for realizing the function of the present invention.

FIG. 8 is a flowchart showing the overall processing flow in the voice input information processing apparatus according to the present invention.

FIG. 9 is a flowchart showing a flow of a setting operation of a mute authentication mode in the voice input information processing apparatus according to the present invention.

FIG. 10 is a flowchart showing a flow of a setting operation of a voiceprint change authentication mode in the voice input information processing apparatus according to the present invention.

FIG. 11 is a flowchart showing a flow of an interrupt routine for performing a mute (change) operation, a voiceprint detection operation, and a voice recognition operation.

FIG. 12 is a conceptual diagram showing an embodiment of a concurrent operation when a timer interrupt is used.

FIG. 13 is a conceptual diagram showing a configuration in a case where a dedicated signal processing means (DSP) is provided separately from a central processing unit (CPU).

FIG. 14 is a conceptual diagram showing an example of a database in which voiceprint feature amounts for user names are recorded.

[Explanation of symbols]

1, 1 ', 1 "voice detecting means, 2, 2', 2" voice generating means, 3 voice signal generating means, 4 voice transmitting means,
5 voice input means, 6 information processing device, 10 sounder,
11 voice input information processing device, 11a central processing unit (C
PU), 11b main storage unit (RAM), 11c auxiliary storage unit (HD), 11d display unit (CRT), 11e keyboard, 11f communication unit, 11g microphone,
11h: A / D converter with amplifier, 11gh, 11'g
h: voice input means, 11i: D / A converter with amplifier,
11j: Speaker, 11ij, 11'ij: Sound output means, 12: Signal processing means (DSP), 13: Voiceprint database.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G10L 15/28

Claims (6)

[Claims] An audio input means having an active mute function for actively silencing the sound by means of a sound detecting means for detecting a sound and a sound generating means for generating a sound having an opposite phase from the detected sound. Thus, the voice input means transmits a signal related to the detected voice to the information processing device and prevents the input voice from leaking to the outside. 2. A voice input means having a voiceprint changing function for actively changing a voiceprint by means of voice detecting means for detecting voice and sound generating means for generating a sound based on the detected voice. The voice input means transmits a signal related to the detected voice to an information processing apparatus, and allows a voice of a different voiceprint to be heard outside. 3. The voice input information processing apparatus according to claim 1, wherein said voice input means performs authentication by extracting voiceprint features of voice. 4. A voice input information processing apparatus comprising: voice detection means for detecting voice; and voiceprint detection means for detecting a voiceprint based on the detected voice. A voice input information processing device, which recognizes only an operation command by a voice of a voiceprint having a characteristic and accepts an operation. 5. The voice input information processing apparatus according to claim 3, further comprising: voice detection means for detecting a voice; and voiceprint detection means for detecting a voiceprint based on the detected voice. A voice input information processing device, wherein the device recognizes only an operation command by voice of a voiceprint having characteristics similar to the characteristics of the authenticated voiceprint, and accepts the operation. 6. The voice input information processing device according to claim 3, wherein at the time of authentication, the voice input information processing device displays or pronounces pronunciation information for authentication, and the authenticator pronounces the pronunciation information, A voice input information processing apparatus for performing authentication on the basis of a voiceprint of a received voice.