JP2001222295A - Individual collating device and recording medium with program for realizing it recorded thereon - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily and inexpensively improve security without increasing remarkable labor and time for adding and registering a hardware in an individual collating device collating whether or not a user is the original. SOLUTION: When the individual identification data are inputted, and an input of identification information is indicated, by using a password after the input is indicated as well as a lapse time from the indication of the input to individual collation, the theft of the password and the assumability by a tape recorder and a voice synthesizer taking times in operation are deterred.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a personal verification device for determining whether or not a user is an identity using an identifier and identification information registered in advance for each user, and a computer program for realizing the same. Related to a recording medium on which is recorded. In particular, even if the password or other identification information is stolen by another person, by checking the elapsed time when the password was actually entered from the password input instruction at the time of personal verification, even if the password was stolen, it could easily be stolen by another person. The present invention relates to a personal collation device for preventing fraud and a recording medium on which a computer program for realizing the same is recorded.

[0002]

2. Description of the Related Art In recent years, with the rapid progress of computers, many people have a certain personal ID and password, such as a cash card, a credit card, and a room check at a company. There is an increasing number of situations where it is necessary to confirm that there is. As the speech recognition technology and the Internet have developed as in today, not only systems such as bank ATMs and entry checks that use ID cards and passwords, but also voice recognition technology to verify whether or not the user is the person himself. Systems and systems for verifying the identity of a person from a remote location via a line have also become widespread, and computer access by telephone (Computer T
The field of elephony integration) and electronic commerce through the Internet and the like are rapidly spreading.

Recently, as systems using voice recognition technology can be realized at low cost, personal verification systems using voice recognition have begun to spread. With speech recognition, passwords can not be misused by simply stealing passwords, and personal verification can be performed from a remote location using a telephone or computer.Therefore, even in the field of computer access by telephone or the field of electronic commerce using the Internet, ,
Personal verification using voice recognition has begun.

[0004] With the development of computer access technology by telephone, important personal information such as home banking balance inquiry and remittance by telephone can be easily operated by inputting a simple personal ID and password. It is becoming possible to pay by entering a simple credit number and password. However, since the method using these lines can be used without a card such as a cash card or a credit card, if identification information such as an identifier serving as a personal ID or a password is known to others, even a telephone or personal computer can be used. There is a problem that if it is used, it can be easily exploited from anywhere, and there is a problem that the person does not notice until the user is abused, and a more secure personal verification device is required.

[0005]

In the case of the conventional personal verification device using voice recognition, the voice information of the person registered in advance is simply verified against the input voice information. When you are speaking, you can record your voice on a tape recorder using a sound-collecting microphone from a distance, even if you are nearby and record that voice on a tape recorder. By doing so, it was possible for another person to spoof using the recorded voice of the person.

When a cash card or the like is required, such as at a bank ATM, even if the password is known, the cash cannot be removed without the cash card itself.
If you steal a cash card, you will often notice it immediately, so it is not easy to misuse it.However, if you operate personal information etc. by phone or personal computer, you can operate it from any phone or personal computer, so it is easy to misuse , There is a lag between the time of the abuse and the discovery of the fact, and the fact that it is possible to operate from any remote location makes it difficult to identify the abused person. .

However, in the case of personal collation using speech recognition, speech synthesis technology has been rapidly advanced along with the advancement of speech recognition technology. In recent years, collecting and analyzing a large amount of voice data could not be realized, but recently collecting a small amount of the voice of the person can analyze the characteristics of the voice and synthesize voice even if it is not collected. It is becoming possible.
In other words, it is becoming possible to synthesize an arbitrary word by resembling the person's voice by recording a few words other than the password in the case of the person's voice.

[0008] For this reason, a technology for discriminating a recorded voice or a synthesized voice from a real voice has been developed. However, this technology needs to be advanced with the progress of the recording technology and the voice synthesis technology. I'm sorry, there is no end. Further, since the discrimination technique is developed in response to a new recording technique and a speech synthesis technique, it follows up and poses a security problem.

[0009] In the case of only inputting a credit number and a password from a personal computer as in electronic commerce on the Internet, if the credit number and the password are stolen, anyone with a personal computer can easily input it from anywhere. E-commerce has progressed since the criminals did not notice that they had been stolen until the invoice was sent. Is a major problem in doing so.

In order to further improve security, it is conceivable to combine technologies such as image recognition.
In the case of identity verification using image recognition, there are variations in verification accuracy due to aging of glasses, beards, hairstyles and faces.
The need to add devices such as image input devices,
There is a problem that it is time-consuming, for example, that it is necessary to take an image in advance. For this reason, there is a need for a technology capable of easily and inexpensively improving security in personal verification without adding hardware or significantly increasing the time and labor required for registration.

[0011]

SUMMARY OF THE INVENTION The present invention relates to a personal verification device for verifying whether or not a user is a user, and can easily and inexpensively provide security without adding a great deal of hardware or adding significant labor when registering. The purpose is to improve. In order to achieve this object, the personal verification device according to the present invention measures the elapsed time from when the user is prompted for voice input to when the voice is actually input, and the elapsed time is measured by the user himself / herself. It is characterized in that it is used as a material for judging whether or not it exists.

[0012] The personal verification device according to the present invention comprises an identifier input unit for inputting an identifier of an individual, an input instruction unit for instructing input of identification information for identifying whether or not the user is an individual, and an elapsed time from the instruction. A time measurement unit for measuring and displaying, the identification information, an identification information input unit for acquiring an elapsed time when the identification information is input from the time measurement unit, the identifier and the identification information in advance, A personal information storage unit for storing information on a time at which identification information is to be input, and a personal verification unit for verifying whether or not the user is a person using the identifier, the identification information, the elapsed time, and the time information. It is characterized by doing.

[0013] With this configuration, it is possible to determine whether or not the input has been made in accordance with the registered time information by using the elapsed time when the identification information is input after the input of the identification information is instructed. Since the time when the identification information is input also functions as a part of the password, security can be easily improved without adding a special device.

Further, the personal verification device according to the present invention is a personal verification device using voice recognition, wherein the identification information input section inputs the identification information by voice, and furthermore, converts the input voice into a predetermined voice. And the identification information stored in the personal information storage unit is audio data converted into the predetermined format.

[0015] With this configuration, it is possible to determine whether or not the voice input has been made in accordance with the registered time by using the time information from when the voice input is instructed until the identification information is voice input. Since the time itself functions as a part of the password, security can be easily improved without adding a special device.

In the personal verification device according to the present invention, in the personal information storage unit, the identification information includes one or a plurality of components, and the time information includes a time to be input for each component. The information is stored, and the identification information input unit inputs the identification information by dividing the identification information for each component, and acquires the elapsed time at the time of inputting the identification information for each part of the identification information.

[0017] With this configuration, the identification information can be divided into a plurality of components and can be input. Since the entire elapsed time when each component of the identification information is input functions as a part of the password, a special password is required. Security can be easily improved without adding devices.

Further, the personal verification apparatus according to the present invention further comprises an identification word storage unit which stores a plurality of words or sentences to be inputted by voice in advance, and wherein the input instruction unit stores the identification word from the identification word storage unit. The personal collation unit has a function of automatically selecting an arbitrary word or sentence and instructing voice input, and determining that the user is not the user unless the time until the voice input is started is within a predetermined time. It is characterized by.

With this configuration, the input identification information is
Since the input is not known until instructed, even if the user attempts to spoof the person using a tape recorder or a voice synthesizer, the voice input will be as much as the operation time of the machine to spoof the person, compared to the user inputting the voice. It takes a long time to
It can prevent others from spoofing.

The computer-readable recording medium on which the program according to the present invention is recorded is a computer-readable recording medium on which a program for causing a computer to execute a step of checking whether or not an individual is an individual is recorded. A step of inputting an identifier of an individual; a step of instructing input of identification information for identifying whether the person is the person; a step of measuring and displaying an elapsed time from the instruction; A step of acquiring an elapsed time at which the identification information is input from a time measurement unit, and a step of storing the identifier and the identification information in advance and information of a time at which the identification information is to be input; Using the identifier, the identification information, the elapsed time, and the time information to check whether or not the user is the principal. .

By loading the recording medium having the above configuration into a computer and executing the same, the user is instructed to input the identification information, and uses the elapsed time at the time when the identification information was input, according to the registered time information. It is possible to judge whether or not the identification information has been input, and the time when the identification information was input itself also functions as a part of the password, so that the personal verification device can easily improve security without adding a special device. Can be realized.

[0022]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The principle and basic operation flow of the present invention will be described with reference to a basic configuration diagram of the present invention shown in FIG. 1 and a flowchart showing a basic processing flow of FIG.

First, the identifier input unit 1 receives a personal identifier input by the user (step S1), and checks whether the input personal identifier is registered in the personal information storage unit 5-1 (step S2). When the personal identifier is not registered in the speaker information storage unit 5-1 (step S2: N
O), the identifier input unit 1 returns to step S1 and waits for input of a personal identifier again.

When the personal identifier is registered in the personal information storage unit 5-1 (step S2: YES), the input instruction unit 2 inputs the identification information such as a password for identifying the user. At the same time as issuing the instruction, an instruction to start time measurement is issued to the time measuring unit 4 (step S3).

When the user is instructed to input the identification information, the user enters necessary identification information with a time interval previously registered as time information in the personal information storage unit 5-1 (step S4). ). And the identification information input unit 3
When the input of the identification information is started, the time when the input of the identification information is started is obtained by referring to the elapsed time measured by the time measuring unit 4 and the elapsed time at the time when the input of the identification information is started Then, the input identification information is passed to the personal verification unit 5 (step S5).

The personal matching section 5 searches the personal information storage section 5-1 based on the personal identifier passed from the identifier input section 1, refers to the time information and the identification information corresponding to the personal identifier, and Is compared with the elapsed time passed from the identification information input unit 3 (step S6), and the result of the comparison is output to another system or the like (step S7).

With this configuration, it is possible to determine whether or not the input of the identification information has been performed in accordance with the registered time information by using the elapsed time when the identification information was input. Since the time when the identification information is input also functions as a part of the password, security can be easily improved without adding a special device.

(Embodiment 1) In the present embodiment, a door opening / closing apparatus in which the present invention is applied to a personal verification apparatus using voice recognition provided on a door that enters and exits a place where security is required will be described as an example. 3 and the flowchart of FIG.

In this embodiment, the input of the personal identifier is performed from the key input device, the input instruction is displayed on the screen of a display device provided adjacent to the key input device, and the input time is set. It is assumed that the elapsed time from is displayed on the screen, and that the identification information is input by voice from a microphone provided next to the key input device. Numeral 7 in FIG. 3 is an input / display device in which a key input device, a display device, and a microphone are integrated into one, and it is assumed that the input / display device is installed on a wall next to the door at a height at which a person can easily input a key. If the user is authenticated as a result of the personal verification, the door automatically opens and closes. If the user is not authenticated, an error message is displayed on the screen.

First, the user inputs a personal identifier such as an employee code from the key input unit of the input / display device 7 (step S11). The personal identifier input by the key is stored in the personal information storage unit 5-1 by the identifier input unit 1.
It is confirmed whether or not it is registered in (step S1
2) If the personal identifier is not registered (step S
12: NO), an error message is displayed on the display unit (step S13), and input of a personal identifier is waited. When a personal identifier is registered (step S12: YE
S) displays a message to input a voice on the display unit of the input / display device 7, and instructs the time measuring unit 4 to measure the elapsed time since the message was displayed (step S14). The personal identifier is passed to the speaker verification unit 5.

When an instruction to input identification information such as a password by voice is displayed on the display unit of the input / display device 7,
The identification information is input by voice in accordance with the time information in the personal information storage unit 5-1 in which how many seconds after the input instruction of the identification information is input beforehand is registered (step S1).
5). In this case, even if a voice is input after a predetermined time after the voice input instruction is displayed, the human time perception is inaccurate.
If the information is displayed on the display unit of the display device 7, the input can be performed accurately. Conversely, for example, time information may be set so as to start voice input after 3 to 6 seconds, and a predetermined time range may be provided to respond to a rough sense of time.

FIG. 5A shows an example of the contents of the personal information stored in the personal information storage unit 5-1. The personal identifier is composed of characters, numbers, symbols, and the like within a predetermined length. In the input / display device 7 of FIG. 3, the key input unit is configured by numbers from 0 to 9, an asterisk symbol, and a sharp symbol, but is not limited thereto.

Next, in the present embodiment, the time information is set in units of seconds. However, this is too long for human beings if the time unit is set to minutes or hours, and is set to milliseconds. Then it is too early. In the case of a system in which a response is made by a machine, the time unit is not particularly limited to the second unit since this is not the case. If you want to start voice input after a predetermined number of seconds, set a specific number as time information, if you want to have a range in the time to start voice input, set the time range as time information Good.

Next, the voice input from the microphone of the input / display device 7 is converted into voice data by the voice converter 3-1. The voice data is obtained by converting the input voice from analog to digital and then performing LPC analysis using LPC analysis.
It is common to use those converted to voice features by the cepstrum method or the MFCC method using FFT analysis, etc., but there is no particular limitation as long as the voice can be verified. An analog-to-digital converted one may be used. In the case of this embodiment,
In the identification information of FIG. 5, audio data converted by any of these methods is stored.

When there is personal information as shown in FIG. 5 (2), when the personal identifier is "# 01", the voice "aiue" is input from the third second after the voice input instruction. When the personal identifier is "# 02", "Kaki" is input from the second second and "Kuke" is input from the fifth second after the voice input instruction. When the personal identifier is "# 20", the voice input of "Sashisuse" is started within 3 to 6 seconds after the voice input instruction. At this time, the voice input of "Sashisetsu" may be started / terminated within 3 to 6 seconds after the voice input instruction. If the personal identifier is "#
In the case of "99", "ra" is set to 5
“Ri” is input by voice at the second, and “Rure” is input by voice at the seventh.
In FIG. 5 (2), the audio data is represented by characters, but actually, the audio data is included in the form of a converted audio feature.

FIG. 6 shows an example in which an input time frame is set for each second and voice is actually input. An input time frame surrounded by a dotted line is a time frame registered in advance as a time for inputting identification information. In the case of the user A, the input time frame of the 4th to 6th seconds is registered as time information, and the voice is input in the input time frame of the 4th to 6th seconds from the input instruction. Becomes OK. On the other hand, in the case of the user B, the input time frame of the third to fifth seconds is registered as time information, but since the voice is input in the input time frame of the second to fourth seconds from the input instruction, Matching based on the elapsed time is NG. In the case of the user C, the input time frame of 5 to 7 seconds is registered as time information, but the time of voice input is slightly shifted. If you want to strictly match time information,
A pattern like user C should be NG,
Actually, since the elderly are slow to respond, the timing of voice input is often slightly shifted. In such a case, it is preferable that the collation of the time information is not performed strictly, and that some deviation is considered. For example, when the input time frame is in units of one second, the verification based on the time information may be OK even if the input time frame protrudes from the time input frame by about 0.2 to 0.3 seconds.

FIG. 5 (3) shows an example in which the elapsed time after the voice input instruction is displayed on the screen so as to be accurately understood. Here, an input time window is provided for each second so that display can be performed for up to 10 seconds. Each time one second elapses, the color of the time frame changes one by one, and the elapsed time is known. In this example, an asterisk is displayed in the input time window for voice input so that the input time frame can be recognized. Because it is understood, it may not be displayed.

For example, if it takes at least 2 seconds to input the identification information and the voice input must be completed within 10 seconds, the start of the voice input must start at the ninth second at the latest. That is, the time during which the voice can be input is the first to ninth seconds. In (2) of FIG. 5, the personal identifier is “# 0
In the case of "1", voice input must be started from the third second, so even if the input voice data is recorded or spoofed using a voice synthesizer, if the time to input the voice data is not known, the spoofing is performed. The probability of success can be reduced to 1/9. In the case where the input of the audio data is divided and the time to be input is set for each, the probability of spoofing can be further reduced.

The voice data registered in the personal information storage unit 5-1 is converted by a voice conversion unit (not shown) in the identification information input unit 3 into analog / digital conversion of the input voice. Is converted into a voice feature by a predetermined method, the voice data is converted by the same method, and the time information when the voice input is started and the converted voice data are passed to the personal matching unit 5 ( Step S16).

The personal verification unit 5 extracts time information and voice data corresponding to the personal identifier from the personal information storage unit 5-1, based on the personal identifier received from the identifier input unit 1. Then, whether or not the time information received from the voice input unit 3 satisfies the time information extracted from the speaker information storage unit 5-1 is checked. The method of voice verification is not particularly limited as long as voice verification is possible. When the elapsed time received from the identification information input unit 3 satisfies the time information extracted from the personal information storage unit 5-1, the personal verification unit 5 also performs verification on voice data (step S17). If the time information is not satisfied or the verification of the voice data does not match (step S18: NO), the verification result indicating the time information mismatch is passed to the door opening / closing unit 6 at that time, and the input / display device 7 displays the result. An error is displayed on the unit (step S19). If the time information is satisfied and the matching of the voice data also matches (step S18: YES),
The door opening / closing unit 6 opens the door 8 for a predetermined time and then closes the door 8 assuming that the user is authenticated (step S2).
0).

Here, in the individual matching, the time information is compared before the voice data is matched because the processing time of the time information is shorter than that of the voice data. . Also, the reason why the voice data is not collated when the time information collation result does not match is that if either the time information or the audio data does not match, it is necessary to consider that the collation does not match. is there. Of course, both collations are performed, and the collation result is transmitted to the door opening / closing unit 6.
Or the order of the time information collation and the audio data collation may be changed.

In this embodiment, a door input / output security system is taken as an example, and a dedicated input / display device is provided. However, when the present invention is used as a computer security system, a key input device may be a keyboard. However, the keyboard is not limited to a hardware keyboard, but may be a software keyboard, a card reader using an ID card, or any other input device that can input a personal identifier. Good.

Further, in the present embodiment, the time information at the time of inputting the voice data from the voice input instruction is used as a part of the collation data, but it is also possible to use the silent time instead. However, when one segment of the voice identification information is long, for example, when a certain segment is “Aieo”, it does not take 5 seconds to pronounce these five characters. In some cases, it may take 4 seconds, and in other cases, the length may not be constant. Therefore, the elapsed time at which the audio data is input is used.

In the present embodiment, the instruction for inputting the identification information and the display of the error are displayed on the screen of the display device. However, they may be simply performed by using indicators such as green and red LEDs. This may be performed using voice or simple sound using a speaker. There is no particular limitation as long as an input instruction and an error display can be made. Further, when the present embodiment is applied to a speaker authentication system using a telephone or the like, the key input device is a push button of the telephone, and the voice input device and the display device are a receiver.

With this configuration, it is possible to determine whether or not the voice has been input at the registered time by using the time information at the time when the identification information was input by voice after the voice input was instructed. Since the time itself functions as a part of the password, the security can be easily improved without adding any special device in the personal verification device using voice recognition.

The time required for inputting the identification information is 1
By setting it longer than seconds, if you do not know the password, use a personal computer etc. to change the password one after another in a short time and enter it, so that it takes a lot of time to break the password, such as searching for the correct password. Has the effect of becoming difficult.

(Embodiment 2) Next, a description will be given of an example in which the personal collating apparatus specifies a word to be input by voice and inputs the word within a predetermined time in Embodiment 1 with reference to FIG. The same parts as those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted.

The difference from the first embodiment at the time of voice input is that the word to be voice input is instructed to the user, so that the impersonator can understand the word to be input, and the input word is divided into a plurality of parts. Therefore, only the first elapsed time when the voice input is started is used as a part of the identification information.

With the advance of personal verification technology using voice recognition and the increase of security and other devices using the technology, those who eavesdrop on identification information that is input by themselves and record it with a tape recorder and attempt to exploit it, Even if it is not possible to record personal identification information directly due to the advancement of speech synthesis technology, if it becomes possible to synthesize speech by collecting only a few voices of the individual, those who try to abuse using a speech synthesis device Is expected to appear. An object of the present embodiment is to prevent abuse by a tape recorder or a speech synthesizer.

First, in the personal identifier 1, when a personal identifier is input from the key input device, the input instructing unit 2 causes a word or a sentence to be voice-inputted as identification information stored in the identifier storage unit 2-1 in advance. Is displayed on the display device together with the message "Please input voice", and the time measuring unit 4 starts measuring time. The user inputs the identification word displayed on the display device by voice. The identification information input unit 3 passes the input voice data and the elapsed time to the personal verification unit 5.

The personal collating unit 5 stores the personal identifier, the time information, and the personal information stored in the personal information storage unit 5-1 based on the personal identifier, the input voice data, and the elapsed time until the voice input is started. Check with audio data. In the case of the present embodiment, since the time until the start of voice input is a predetermined time common to all users, the time information in the personal information storage unit may be all the same, or The information may not be provided in the information storage unit 5-1, but may be provided in the personal verification unit 5. The voice data needs to have all the patterns of the identification information registered in the identification information storage unit 2-1.

In the case of this embodiment, how long the time information is within is determined by the time required for the user to input a specified word or sentence by voice and the time required for operating the tape recorder or voice synthesizer. It is necessary to decide on a balance. This is because when someone tries to spoof using a tape recorder or a speech synthesizer, it is not possible to know in advance which identification information to enter among a plurality of pieces of identification information. Until the identification information is input, the time required for the impostor to operate the tape recorder or to input the identification information to be synthesized into the voice synthesizer from the keyboard is longer than when the person himself inputs the voice. This is because such a thing is used.

In the case of a tape recorder, it takes considerably longer to operate a tape to extract necessary voices than in the case where the user himself inputs voices. In the case of a voice synthesizer, identification information is input from a keyboard. Requires three to four times as much as voice input (Electronic and Information Expensive Introduction Series 2 Sound and Speech Engineering Chapter 14 Speech Recognition 174
Page Sadahiro Furui, Modern Science Company, September 1992). Therefore, if the time until the start of voice input is set to a time that is sufficiently shorter than the time for operating the tape recorder or the voice synthesizer, if the verification of the voices input within the predetermined time matches, the user will be notified. You can determine that you are the person.

At this time, the predetermined time is sufficiently shorter than the time required for operating the tape recorder or the voice synthesizing device, and is set to be within 2 seconds or 3 seconds as a time sufficient for the user to start voice input. Is preferred, but it is not particularly limited to this time. This is because the operation time of the impostor may vary depending on the length of words or sentences to be input or the number of registrations. In other words, if the identification information to be input by voice is too short, the time for keyboard input to the voice synthesizing apparatus can be reduced, and there is not much time difference from the case where the user inputs voice, so it takes much longer than inputting voice. It is preferable to indicate identification information having such a length. In addition, if the number of words or sentences to be input is small, the position of the tape recorder can be shortened. Therefore, it is preferable that the number of registrations is not very small. Also, if a single voice input instruction is issued, if the voice of the identification information corresponding to the tape portion that happens to be positioned by the tape recorder is included, the verification will be OK. Repeat multiple times until all voices are input within the specified time.
Finally, the collation may not be made OK.

With this configuration, the identification information to be input is not known until the input is instructed. Therefore, even if another person attempts to spoof using a tape recorder or a speech synthesizer, the identification information is not compared with the case where the user himself inputs the voice. The operation time is longer, and it is possible to prevent others from spoofing.

(Embodiment 3) Next, in Embodiment 2, the personal verification device specifies a word having a different number of syllables and repeats the input of the word a plurality of times. An example in which individual verification is performed using fluctuations will be described with reference to FIG. 7 that is the same as in the second embodiment, focusing on differences from the second embodiment.

The major difference from the second embodiment at the time of voice input is that the voice input of a word having a different number of syllables is repeated a plurality of times to perform collation, and the elapsed time is not the elapsed time itself but the difference between the elapsed times. The point is that the fluctuation of the time difference is used for collation.

This embodiment differs from the first embodiment in that the time for positioning the tape is different when spoofing is performed using a tape recorder, and that when using a speech synthesizer to spoof a word, it takes more time to input a keyboard with longer words. Focusing on the point, an object is to suppress spoofing using a tape recorder or a speech synthesizer.

First, in the personal identifier 1, when the personal identifier is input from the key input device, the input instructing unit 2 reads out from the identifier storage unit 2-1 in which a plurality of words having different numbers of syllables are registered in advance. An arbitrary one is selected and displayed on the display device together with the message “Please input voice”. The user inputs the words displayed on the display device by voice. This is repeated several times while selecting words with different syllables. The identification information input unit 3 passes all the input voice data and the elapsed time to the personal verification unit 5.

The personal collation unit 5 compares the personal identifier and the voice data stored in the personal information storage unit 5-1 with each other based on the personal identifier and the input voice data and the elapsed time until the voice input is started. The matching is performed, and the difference between the elapsed times is obtained.

In the case of the present embodiment, the collation is performed not only by using the voice but also by using the fluctuation of the time difference. The fluctuation of the time difference can be set for each individual. Value. Therefore, all users are 1
One predetermined value may be used, and the time information in the personal information storage unit may be all the same, or the personal information storage unit 5-1 may have one time information but not the time information in the personal information storage unit 5-1. The audio data needs to have all the patterns of the identification information stored in the identification information storage unit 2-1.

The number of seconds within which the time information of the fluctuation of the time difference is set may vary depending on the number of syllables of a word indicating a voice input. After the words indicating voice input are displayed,
It is considered that the time required for recognizing the same is almost the same when the user himself inputs a voice and when he tries to spoof using a tape recorder or a voice synthesizer. However, there is not much difference in the time until the person himself starts voice input of the specified word, depending on the number of syllables of the specified word.
In the case of spoofing using a tape recorder, the time required for positioning the tape varies depending on the position of the word to be input. In the case of spoofing using a speech synthesizer, words with longer syllables use speech synthesis from the keyboard. It takes time to enter the word you want. In other words, it is necessary to designate a word having a difference in the number of syllables so that the time difference until the speech input becomes clear. For example, it is preferable to indicate a word having a sufficient difference in the number of syllables, such as a word of three syllables and a word of ten syllables.

For example, when a three-syllable word and a five-syllable word are specified, and when a three-syllable word and a ten-syllable word are specified, the difference in the number of syllables is two syllables in the former and seven syllables in the latter. For this reason, when inputting via a keyboard in a speech synthesizer, the latter obviously takes more time, and the fluctuation of the time difference becomes larger. Therefore, the matching accuracy is higher when a word having a large difference in the number of syllables is specified.

There is an individual difference in the time until the designated word is recognized, and a time difference occurs even in the same individual.
In order to reduce the probability of recognizing that a person is not the person himself, it is necessary to allow time for recognizing the designated word with a margin. That is, it is necessary to make the permissible range of the fluctuation of the time difference somewhat long. Then, in the case of the second embodiment, if the instructed word is short, the time required for keyboard input to the speech synthesizer is also short, and the probability that the impostor is erroneously recognized as the person is increased. Therefore, it is preferable that the difference in the number of syllables of the designated word is large.

As described above, it is possible to input words having different numbers of syllables by voice and use the fluctuation of the time difference of the elapsed time until the voice input to help identify the user. Even if the difference in the number of syllables is large, mechanical operation is performed by unifying the elapsed time before inputting a voice to the time that can be input even for a long word (for example, in a speech synthesizer, However, if it is possible to input after 8 seconds, regardless of the length of the syllable, the speech is unconditionally synthesized after 8 seconds). It is more effective to start the voice input in time.

According to this configuration, when a word having a different number of syllables is designated and voice input is performed, the fluctuation of the time until the input is reduced when the user inputs the voice, whereas the number of syllables is longer when the voice is synthesized using the voice synthesizer. The longer the time until the voice input takes, the greater the fluctuation of the time difference until the input, and if a fake using a tape recorder, the time difference occurs in the positioning operation depending on where the corresponding voice is recorded on the tape, By judging the fluctuation of the time difference, it is possible to prevent others from spoofing.

(Embodiment 4) Next, unlike Embodiments 1 and 2, an example of an embodiment in which identification information is input by normal key input without using voice recognition will be described with reference to the configuration diagram of FIG. Will be explained. As a typical example, an example in which cash is withdrawn from an ATM when the present invention is incorporated in an ATM of a bank will be described. The same parts as those of the first embodiment are denoted by the same reference numerals, and description thereof is omitted.

First, "payout" is selected from the menu displayed on the input / output screen 9 using the ATM transparent touch panel. Then, a message “Please insert a cash card” is displayed on the input / output screen 9. When a cash card is inserted into the ATM card reading unit 10, the input instruction unit 2 displays a key for inputting a password, a message "Please enter a password", and a software keyboard for input on the input / output screen 9. At the same time, it instructs the time measuring unit 4 to measure the passage of time. At this time, the elapsed time may be displayed on the input / output screen 9 as in (3) of FIG.

Next, the user inputs a password as identification information in accordance with time information registered in personal information storage unit 5-1. When the displayed key is pressed with a finger, the coordinates of the pressed position are input, and the identification information input unit 3
Converts the input coordinates to character data displayed at the position where the finger is pressed, regards that character as being input, and calculates the input character data and the elapsed time at the time of inputting it as a time measurement unit. 4 and pass it to the personal verification unit 5.

The personal information storage unit 5-1 stores character data as a password instead of the voice data of the first embodiment. The personal collation unit 5 collates the personal information stored in the personal information storage unit 5-1 based on the personal identifier recorded on the cash card, the character data input by the input instruction unit 2, and the elapsed time. If the collation results do not match, the process returns to the password input again. More than a predetermined number of times,
If the verification error continues, the display returns to the menu. If the collation results match, the process proceeds to the amount input unit 11, where the input screen for the amount to be paid out is displayed on the input / output screen 9, and the amount to be paid out is entered. If the input amount is a payable amount, the input amount is paid out in the amount paying unit 12.
If the entered amount cannot be paid out due to insufficient balance, an error is displayed and the amount is entered again.

In the present embodiment, a dedicated input device and output device are provided in order to take cash as an example by a bank ATM, but the input device and the output device for inputting a personal identifier and a password are provided. The device is not particularly limited as long as the device can display an input instruction or an error, and can operate according to the collation result.

Although the input of the identification information and the display of the error are displayed on the screen of the display device in the present embodiment, they may be simply performed by an indicator using green or red LEDs or the like. This may be performed using voice or simple sound using a speaker. Further, when the present embodiment is applied to a personal verification system using a telephone or the like, the key input device becomes a push button of the telephone, and the voice input device and the display device become a handset.

Further, as in the second embodiment, a password as identification information may be divided into a plurality of components.

With this configuration, it is possible to determine whether or not the input of the identification information has been performed according to the registered time by using the time information from when the input of the identification information is instructed until the time when the identification information is actually input. In addition, since the time of voice input itself also functions as a part of the password, security can be easily improved without adding a special device.

(Embodiment 5) Next, a recording medium in which a program for realizing the personal verification device according to the embodiment of the present invention is described, as shown in FIG.
Not only the portable storage medium 13 such as the ROM 13-1 and the floppy disk 13-2, but also any other storage device 14 stored at the end of a communication line, or the storage medium 15 such as a hard disk or a RAM of a computer may be used. When the personal verification device according to the present invention is used, it is read by the computer 16 and executed on the main memory.

By loading the program from the recording medium into the computer, it is possible to check whether or not the identification information has been input at the registered time using the time information from when the input of the identification information was instructed to when the identification information was actually input. It is possible to determine whether or not the personal information verification device can improve the security easily without adding a special device because the time when the identification information is input itself also functions as a part of the password. become.

[0077]

According to the personal verification apparatus of the present invention, whether or not the input of the identification information has been performed according to the registered time information by using the elapsed time at the time when the identification information was input is instructed. Can be determined, and the time itself when the identification information is input also functions as a part of the identification information. Therefore, security can be easily improved without adding a special device.

Further, according to the personal verification device using voice recognition according to the present invention, since identification information to be input is not known until an input is instructed, another person may attempt to spoof using a tape recorder or a voice synthesis device. However, the operation time of the machine is longer than when the person inputs voice, and it is possible to prevent others from spoofing.

Further, according to the personal verification apparatus using voice recognition according to the present invention, when a word having a different number of syllables is specified and voice input is performed, the fluctuation of the time until the input is small when the user inputs voice. When using a speech synthesizer, the longer the number of syllables to be input, the longer it takes to input the voice, the greater the fluctuation of the time difference until the input, and the more falsely using a tape recorder, There is a time difference in the positioning operation depending on whether or not the sound to be recorded is recorded, and by determining the fluctuation of the time difference, it is possible to prevent others from spoofing.

[Brief description of the drawings]

FIG. 1 is a basic configuration diagram of the present invention.

FIG. 2 is a flowchart showing a basic processing flow of the present invention.

FIG. 3 is a configuration diagram of a personal verification device according to the first embodiment of the present invention.

FIG. 4 is a flowchart showing the flow of the first embodiment of the present invention.

FIG. 5 Examples of various information

FIG. 6 shows an example of voice input in which an input time frame is set.

FIG. 7 is a configuration diagram of a personal verification device according to a second embodiment of the present invention.

FIG. 8 is a configuration diagram of a personal verification device according to a fourth embodiment of the present invention.

FIG. 9 shows an example of a storage medium.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Identifier input part 2 Input instruction part 2-1 Identification word storage part 3 Identification information input part 4 Time measurement part 5 Personal collation part 5-1 Personal information storage part 6 Door opening and closing part 7 Input display device 8 Door 9 ATM input / output Screen 10 ATM card reading unit 11 Amount input unit 12 Amount payment unit 13 Portable storage medium such as CD-ROM or floppy disk 13-1 CD-ROM 13-2 Floppy disk 14 Other storage stored at the end of the communication line Device 15 Storage medium such as RAM / hard disk on computer 16 Computer

Claims (5)

[Claims] A personal identification device for collating whether or not the individual is an individual; an identifier input unit for inputting an identifier of the individual; and an input for instructing input of identification information for identifying whether or not the individual is an individual. An instructing unit, a time measuring unit that measures and displays an elapsed time from the instruction, the identification information, and an identification information input unit that acquires an elapsed time when the identification information is input from the time measuring unit, A personal information storage unit for storing in advance the identifier, the identification information, and information on the time at which the identification information is to be input; and a person using the identifier, the identification information, the elapsed time, and the time information. A personal collation device comprising a personal collation unit for collating whether or not the personal collation is performed. 2. The personal verification device according to claim 1, wherein the personal verification device is a personal verification device using voice recognition, wherein the identification information input unit inputs the identification information by voice,
Furthermore, the personal information storage unit has a function of converting the input voice into voice data of a predetermined format, and the identification information stored in the personal information storage unit includes:
The personal verification device, wherein the personal data is audio data converted into the predetermined format. 3. The personal verification device according to claim 1, wherein, in the personal information storage unit, the identification information includes one or a plurality of components, and the time information includes at least one component. The identification information input unit inputs the identification information by dividing the identification information for each component, and obtains the elapsed time at the time when the identification information is input for each part. A personal verification device characterized by the following. 4. The personal verification device according to claim 2, further comprising: an identification word storage unit that previously stores a plurality of words or sentences to be input by speech, wherein the input instruction unit stores the identification word. The unit automatically selects an arbitrary word or sentence from the unit and instructs voice input, and the personal verification unit has a function of determining that the user is not the user unless the time until the start of the voice input is within a predetermined time. A personal verification device, comprising: 5. A computer-readable recording medium storing a program for causing a computer to execute a step of checking whether or not an individual is an individual, comprising the steps of: inputting an identifier of the individual; Instructing input of identification information for identifying whether or not, measuring and displaying an elapsed time from the instruction, the identification information, and an elapsed time when the identification information is input from the time measuring unit. Obtaining the identifier, the identifier and the identification information in advance, and storing information of the time when the identification information is to be input; and the identifier, the identification information, the elapsed time, and the time information. A computer-readable recording medium for storing a program to be executed by a computer, the method including a step of checking whether or not the user is the user.