JP2005091991A - Personal identification system, personal identification method, and storage device used for personal identification system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a personal identification system having a high recognition rate that can reduce the memory capacity of a storage device, can reduce the cost of the storage device, and is less affected by the state of a person to be identified.
A portable storage device that stores in advance a first time interval between a plurality of predetermined characters, a plurality of characters, and a predetermined frequency from a voice uttered by a specific individual for a predetermined series of characters, and a storage device The first time interval stored, the reading means for reading a plurality of characters and frequencies, the voice input means for inputting the voice uttered by the identified person, and the first time interval from the plurality of voice characters inputted by the voice input means Audio data extraction means for extracting a corresponding second time interval and a frequency corresponding to a predetermined frequency, and the second time interval and frequency extracted by the audio data extraction means and the first time interval and frequency read by the reading means And an information processing apparatus having collation means for collating them.
[Selection] Figure 1

Description

  The present invention relates to a personal identification system, a personal identification method, and a storage device used in a personal identification system for identifying a specific individual by collating with information on the specific person stored in advance in the storage device.

  Conventionally, there has been an information processing apparatus described in JP-A-5-174203 as a method for identifying a specific individual. This device receives application of a specific service by voice recognition using information in the IC card, and is composed of a voice input device and an IC card that can be attached to and detached from the voice input device. This IC card has collation means for determining whether or not to use the information in addition to specific personal information and unique information.

  As another example, there is a personal identification mechanism described in Japanese Patent Laid-Open No. 6-215017. In this identification mechanism, the voice signal input from the voice input mechanism is compared and identified with the voice data of the user registered in advance, and the voice signal of the user input from the voice input mechanism is analyzed and recognized. Whether or not the user is the user is identified by comparing with a keyword registered by being converted into a character code as a keyword.

Another example is a personal identification system described in Japanese Patent Laid-Open No. 5-18147. In this method, the card carrier can check the physical or chemical characteristics obtained from the card holder against data registered in advance in an IC memory embedded in the card. It was used to identify whether or not a specific person is registered in the memory in the card.
JP-A-5-174203 Japanese Patent Laid-Open No. 6-215017 Japanese Patent Laid-Open No. 5-18147

  In the information processing apparatus described in the above-mentioned Japanese Patent Application Laid-Open No. 5-174203, since the collation means is stored in the IC card in addition to the specific personal information and the unique information, the storage capacity must be increased. The cost of the card was high.

  Further, in the personal identification mechanism described in Japanese Patent Application Laid-Open No. 6-215017, as the data for personal identification, the user's own voice data and a keyword obtained by converting the voice signal of the user into a character code are required. In order to store these voice data and keywords, a storage medium having a large storage capacity and high cost is required.

  Furthermore, in the personal identification system described in Japanese Patent Laid-Open No. 5-18147, the physical or chemical characteristics change due to the cardholder's state, so that the identification rate of a specific individual varies greatly. There was a fear. In addition, data relating to physical or chemical characteristics requires a large storage capacity even when compressed, and the cost of the card is high.

  In order to solve the above problems, in the personal identification system of the present invention, a first time interval between a plurality of predetermined characters and a plurality of characters are stored in advance from a voice uttered by a specific individual for a predetermined series of characters. Portable storage device, first time interval stored in the storage device and reading means for reading a plurality of characters, voice input means for inputting a voice uttered by the identified person, a plurality of voices input by the voice input means Voice data extracting means for extracting a second time interval corresponding to the first time interval from the character of the character, and collating the second time interval extracted by the voice data extracting means with the first time interval read by the reading means And an information processing apparatus having collation means.

  In the personal identification system of the present invention, a mobile phone in which a first time interval between a plurality of predetermined characters, a plurality of characters, and a predetermined first frequency are stored in advance from a voice uttered by a specific individual for a predetermined series of characters. Possible storage device, first time interval stored in the storage device, a plurality of characters, a reading means for reading the first frequency, a voice input means for inputting a voice uttered by the identified person, and a voice input means Voice data extracting means for extracting a second time interval corresponding to the first time interval and a second frequency corresponding to the first frequency from the plurality of characters of the recorded voice, and a second extracted by the voice data extracting means And an information processing apparatus having a collating unit that collates the first time interval and the first frequency read by the reading unit with the time interval and the second frequency.

In the personal identification method of the present invention, a step of storing a predetermined first time interval, a plurality of characters, and a predetermined first frequency in a portable storage device from a voice uttered by a specific individual for a predetermined series of characters. Reading a first time interval stored in the storage device, a plurality of characters, and a predetermined first frequency, inputting a voice uttered by the identified person,
Extracting a second time interval corresponding to the first time interval and a second frequency corresponding to the first frequency from a plurality of characters of the voice input by the voice input means; and a second extracted by the voice data extraction means. Collating the time interval and the second frequency with the first time interval and the first frequency read by the reading means;
It is characterized by having.

  Preferably, the first time interval between the predetermined plurality of characters is a time interval of speech of the first and last characters in the predetermined series of characters.

  The first frequency is preferably the highest frequency among the predetermined formant frequencies possessed by the voice uttered by a specific individual for a predetermined series of characters.

  It is preferable that the utterance of a specific individual for a predetermined series of characters is performed a plurality of times, and the average value and standard deviation of the first time interval or the first frequency in each utterance are stored in the storage device.

  The storage device is preferably an RF-ID card, an RF-ID tag, or a contact IC card.

  The storage device and the information processing device preferably have means for exchanging information wirelessly.

  It is preferable that the storage device includes an extraction storage device that stores the first time interval and the first frequency.

  The storage device used in the personal identification system of the present invention preferably has means for generating drive power by electromagnetic induction.

  According to the present invention, since the card holder's unique information and personal information extracted by a predetermined process are stored in the storage device, the memory capacity can be reduced and the cost of the storage device can be reduced. Can be suppressed. In addition, since a specific individual is identified using a time interval between a plurality of authentication characters and a frequency selected according to a predetermined criterion, the recognition rate can be increased. Furthermore, since the cardholder is determined to be a cardholder if the time interval and the predetermined frequency of the authentication character by the utterance of the identified person are within the predetermined range, the influence of the state of the identified person is reduced.

  Hereinafter, an example in which an embodiment according to the present invention is used for security when entering and leaving a room will be described in detail with reference to the drawings. The present invention is not limited to security at the time of entering and leaving a room, but can also be used for identifying a computer user, for example. The identification system according to the present embodiment relates to a system for identifying an individual based on information stored in a portable RF-ID card 10 (FIG. 1), and includes an RF-ID card 10 and an RF-ID. A data extraction device 20 (FIG. 2) that extracts and stores information on the person to be identified in the card, and reads the information stored in the RF-ID card and compares it with the input data to determine whether or not it is the card owner. It has the information processing apparatus 40 (FIG. 1) to collate.

  As shown in FIG. 1, the RF-ID card (storage device) 10 includes personal information storage means 11 for storing personal information of the card holder (specific individual) and unique information (biological information) of the card owner. ) And the information processing apparatus provided outside, and the storage and reading of information in the personal information storage unit 11 and the specific information storage unit 12 are controlled, and electromagnetic waves are received from received radio waves. It has communication / control / power generation means 13 for generating power to be driven by induction, and a transmission / reception antenna 14. The personal information storage means 11 stores personal information such as the cardholder's name, gender, date of birth, and registration code number. The unique information storage means 12 stores unique information (biological information) based on the voice as information unique to the cardholder's biometric. The communication / control / power generation means 13 and the transmission / reception antenna 14 are provided in an existing RF-ID card. By providing these, the information processing apparatus and information arranged at a distance of several mm to several m are provided. Can be exchanged. In order to store information in the personal information storage means 11 and the unique information storage means 12, the RF-ID card 10 is connected to the data extraction device 20 shown in FIG. In the present embodiment, an RF-ID card having a wireless function is used as a storage medium. However, other storage media such as a contact IC card can be used, and it is connected to the information processing apparatus 40 instead of wireless. Information may be exchanged.

  As shown in FIG. 2, the data extraction device 20 includes a microphone (speech input means) 21 for inputting a sound produced by a cardholder uttering a predetermined series of characters, and a sound input from the microphone 21. An input processing unit 22 that converts the current value into a binary value, a voice CODEC unit 23 that converts voice data obtained by the input processing unit 22 into character data, a timing processing unit 24, and an authentication character extraction Unit 25, timing data extraction unit 26, and audio frequency component extraction unit 27. The timing processing unit 24 detects the start and end timing of the voice data of each character issued by the cardholder from the data obtained by the input processing unit 22 and the voice CODEC unit 23. The authentication character extraction unit 25 extracts data corresponding to a plurality of characters necessary for authentication from the data output from the voice CODEC unit 23 based on the timing obtained by the timing processing unit 24. The timing data extraction unit 26 extracts time intervals (first time intervals) between a plurality of authentication characters necessary for authentication from the timing obtained by the timing processing unit 24 based on the extraction result in the authentication character extraction unit 25. To do. The voice frequency component extraction unit 27 calculates a frequency corresponding to the voice data of the target character based on the timing obtained by the timing processing unit 24, and extracts one frequency (first frequency) according to a predetermined condition. To do. This extraction method will be described later.

  Here, a plurality of characters to be extracted by the authentication character extraction unit 25 are set in advance using the authentication character setting means 29 connected to the timing processing unit 24. For example, the first and last characters of the input voice are input from input means (not shown) such as a keyboard provided in the authentication character setting means 29. In the voice CODEC unit 23, it is also possible to designate a character that can easily convert voice into a character on the data extraction device side and set it as an authentication character.

  The frequency extracted by the audio frequency component extraction unit 27 is set in advance using the frequency setting unit 30 connected to the audio frequency component extraction unit 27. For example, the setting of the frequency will be described below assuming that the voice frequency component extraction unit 27 extracts the voice data having the highest frequency among the frequencies corresponding to each character. It is also possible to prepare a plurality of setting methods and select from among them. In this case, the selection method is preferably stored in the unique information storage means 12 of the RF-ID card 10.

  The sound is input a plurality of times, and the arithmetic unit 31 calculates an average value and a standard deviation for the time interval (second time interval) extracted by the timing data extraction unit 26 for each input sound. The voice frequency component extraction unit 27 extracts the voice frequency (second frequency) a plurality of times, and the calculation unit 31 measures the average value and standard deviation of the highest frequency of the voice. Note that a numerical value (for example, variance) other than the standard deviation may be used as long as it indicates a variation in the number of measurements.

  In this way, a predetermined series of characters, a plurality of authentication characters, the average value and standard deviation of the calculated time intervals between the plurality of authentication characters, and the average value and standard deviation of the highest frequency of voice are Is stored in the specific information storage means 12 of the RF-ID card 10 connected to the output unit 32 in a contact or non-contact manner at the time of storage. The information stored in the unique information storage unit 12 is held in the unique information storage unit 12 including the nonvolatile memory of the RF-ID card 10 even after the RF-ID card 10 is removed from the output unit 32. As described above, since the average value and standard deviation of the time intervals between the plurality of authentication characters and the average value and standard deviation of the highest frequency of the voice are stored in the unique information storage means 12, the storage capacity is extremely high. A small one is enough. If a predetermined series of characters is not required for the cardholder's check, it is not necessary to store them in the unique information storage means 12, so that the storage capacity can be further reduced. Furthermore, the storage capacity can be further reduced by limiting the information to be stored to the average value of the time interval and the average value of the audio frequency. In this case, a predetermined value is separately set instead of the standard deviation. The personal information is stored in the personal information storage means 11 by an existing tag storage device.

  As shown in FIG. 1, the information processing apparatus 40 includes a communication / control unit 42, an audio data extraction unit 45, and a collation unit 46. The communication / control means 42 can communicate with the RF-ID card 10 via the transmission / reception antenna 41, and the personal information storage means 11 and the unique information by the personal information / unique information confirmation means (reading means) 43 connected thereto. The reading of information stored in the information storage means 12 is controlled. The voice data extracting unit 45 is configured to extract a time interval between a plurality of authentication characters and a voice from a voice input from a microphone (voice input unit) 44 by a person carrying the RF-ID card 10 (identified person). Extract the highest frequency. The collation unit 46 collates the stored personal information with the information stored in the personal information storage unit 11, and the unique information stored in the unique information storage unit 12 and the voice data extraction unit 45 extract it. Check against information. Here, the voice input from the microphone 44 is a voice produced by uttering a predetermined series of characters identical to those input from the microphone 21 at the time of registration of the unique information.

  When the collation means 46 determines that the person carrying the RF-ID card 10 is the owner of the RF-ID card 10 as a result of collation, the collation means 46 opens the door to the electric lock 50 at the entrance / exit connected thereto. A control signal is output, and the electric lock 50 is automatically opened in response to this signal. If it is determined that the person carrying the RF-ID card 10 is not the owner of the RF-ID card 10 as a result of the collation, the collating means 46 does not output any control signal to the electric lock 50 and the electric lock 50 Keeps closed.

  Thus, since it identifies using the time interval between several authentication characters, and the highest frequency of an audio | voice, it is less influenced by a card | curd carrier's state and can identify with high probability. Further, since the information necessary for identification stored in the RF-IF card 10 includes a plurality of authentication characters, a time interval between the authentication characters, and the maximum frequency of the voice, the memory storage capacity can be reduced. Can be reduced. Furthermore, since a separate power source is not required, the card to be carried can be reduced in size and thickness.

Subsequently, an identification operation using the identification system of the present embodiment will be described.
First, before performing the identification operation, personal information and unique information of the cardholder are stored in advance in the RF-ID card 10 using the data extraction device 20. In this embodiment, “hiratake sesame” is selected as a series of characters to be uttered by the cardholder, and the first “hi” and the last “ma” are set as authentication characters by the authentication character setting means 29. Further, the highest frequency is set by the frequency setting unit 30 as the frequency to be extracted. As a series of characters to be uttered by the cardholder, any character can be selected as long as it is two or more characters. In addition, as the authentication character, an arbitrary word at an arbitrary position in a series of characters can be set, and an arbitrary number of characters can be set as long as the number of authentication characters is two or more. Further, in addition to the highest frequency, for example, the lowest frequency and the average frequency can be selected as the frequency to be extracted, and the extraction range may be, for example, only the authentication characters in addition to all the spoken characters.

  When the voice of “Hirake Sesame” uttered by the cardholder is input from the microphone 21, the timing processing unit 24 calculates the timings of “Hi”, “Ra”, “Ke”, “Go”, and “Ma”, The timing data extraction unit 26 extracts the time interval between the first “hi” and the last “ma” as authentication characters. This time interval is the interval between the utterance start timings of the authentication characters “hi” and “ma”, but of course may be the interval between the utterance end timings, or may be the interval between the peak values of the utterance volume. The time interval data is stored in the calculation unit 31.

  On the other hand, in the extraction of the audio frequency, the audio frequency component extraction unit 27 extracts and compares the frequencies corresponding to the respective audio data of “hi”, “ra”, “ke”, “go”, and “ma” and finds the highest frequency. Do by selecting. The frequency data is stored in the calculation unit 31.

  Such extraction work is repeated a plurality of times (for example, 5 times), and the arithmetic unit 31 calculates an average value and a standard deviation for each of the time interval data and the frequency data, and outputs the result from the output unit 32, It is stored in the unique information storage means 12 together with a series of characters and authentication characters.

Next, an identification operation when entering a room where the identification system of the present embodiment is introduced will be described with reference to FIG.
When a person carrying the RF-ID card 10 in which personal information and unique information are stored in advance enters the room as described above, the RF-ID card 10 is held over the information processing device 40 arranged near the electric lock 50 ( Step S101). Then, the information processing apparatus 40 that has been in a standby state becomes an operating state, and communication is performed between the transmission / reception antenna 41 and the transmission / reception antenna 14.

  Subsequently, the card holder utters a predetermined series of characters to the microphone 44 and inputs the voice (step S102). Next, a match between the set of characters and the voice generated by the cardholder is checked (step S103). If the uttered character is different from the set of set characters, when the voice data extracting means 45 determines, the subsequent processing is not performed and the identification ends (NO in step S103).

  When it is determined that the uttered character is a series of characters read from the card (YES in step S103), the voice data extraction means 45 uses the authentication characters read from the card (in this embodiment, “hi” and “ M ") is extracted and these time intervals are measured, and the highest frequency is extracted from the voice data of each character (step S104).

  Note that step S103 is omitted, and after inputting the voice (step S102), two predetermined characters are extracted without checking the match between the set of characters and the voice generated by the cardholder. It is also possible to measure the time interval between them and extract the audio frequency component data (step S104). In this case, the authentication data is extracted from the voice generated by the card holder by the voice data extracting means 45, and these time intervals are measured. If this extraction is not possible, the identification is terminated without performing subsequent processing.

  On the other hand, in the personal information / unique information confirmation unit 43, the average value of the time interval, the standard deviation, and the average value of the highest frequency of the voice stored in the unique information storage unit 12 via the transmission / reception antenna 41 and the communication / control unit 42. And the unique information consisting of the standard deviation is read (step S105). The read unique information is collated by the collating means 46 with the time interval data and the highest frequency data extracted from the voice uttered by the card holder (step S106).

The collation is performed in the collation means 46 as follows.
First, the difference between the time interval and the maximum frequency extracted from the voice uttered by the card holder and the average of the time interval and the maximum frequency stored in the unique information storage means 12 is obtained. Subsequently, the absolute value of these differences is multiplied by 10 and divided by the respective standard deviations stored in the unique information storage means 12. When 50 is added to the value thus obtained, a deviation value for the voice data stored in the specific information storage means 12 of the time interval and frequency extracted from the voice uttered by the cardholder can be obtained. If this deviation value is 45 or more and 55 or less, the time interval and the maximum frequency extracted from the voice uttered by the card holder are close to the average time interval and the average maximum frequency stored in the unique information storage means 12, respectively. As a matter of fact, it is determined that the card holder is the card holder. This deviation value range can be arbitrarily set according to the required security state.

A more detailed description will be given with reference to FIGS. 4 and 5 regarding collation of the highest frequency extracted from the voice.
When words are generated as speech, the speech is composed of vowels and consonants. While the consonant does not have a stable frequency spectrum of peaks, the frequency spectrum of vowels has a relatively stable peak. This peak is generally called formant frequency. Since the formant frequency depends on the shape of the human trachea, it shows different values depending on the individual. In this embodiment, an individual is identified using a formant frequency having such characteristics.

FIG. 4 is a block diagram showing a detailed configuration of the voice data extraction unit 45 and the collation unit 46 shown in FIG. As shown in FIG. 4, the audio data extraction means 45 includes an audio CODEC unit 51, an FFT or autocorrelation function calculation circuit 52, and a storage device 53.
The matching means 46 includes a target frequency component setting circuit 61, a selection circuit 62, and a comparison circuit 63.
Have

  In FIG. 4, the analog audio signal input from the microphone 44 is converted into digital audio data by the audio CODEC unit 51, and the frequency spectrum is obtained from the audio data by the continuous FFT or autocorrelation function in the FFT or autocorrelation function calculation circuit 52. It is done. In this frequency spectrum analysis, a stable vowel frequency peak can be obtained with time. For each vowel of the first vowel to the fifth vowel, the first formant frequency, the second formant frequency, and the third formant frequency, The third formant frequency is stored in the storage device 53.

  Here, FIG. 5 shows the vowels “a” (a), “i” (i), “u” (u), “e” (e), and “o” (o) from the lower frequency side. The average value of the peak (1st formant frequency, 2nd formant frequency, and 3rd formant frequency) of the frequency spectrum to the 3rd is illustrated.

  The frequency to be collated is selected from the formant frequencies stored in the storage device 53 by the selection circuit 62 according to the target frequency component setting circuit 61, and the frequency data is output to the comparison circuit 63. In the present embodiment, the maximum frequency is selected from the first formant frequency region, but the maximum frequency can also be selected from the second formant frequency region or the third formant frequency region.

  On the other hand, the comparison circuit 63 reads the individual specific formant frequency data (personal information data) registered in advance in the RF-ID card 10 from the personal information / unique information confirmation means 43 and selects the selection input from the selection circuit 62. Compared with the frequency data, if they match, a collation output is generated.

  If it is determined by the above verification that the cardholder is the cardholder (YES in step S107), if the verification is performed by verification of personal information (not shown), the verification means 46 controls the electric lock 50. A signal is output and the electric lock 50 is automatically opened. On the other hand, if it is determined that the cardholder is not the cardholder (NO in step S107), the voice is input again and the verification operation is performed. If the card carrier is not determined to be the cardholder even after a predetermined number of times of collation by providing a counting unit in the collating means 46, voice input from the microphone 44 is performed thereafter. It is preferable not to do so because it is possible to eliminate the case where the card owner is determined by inputting the voice very many times.

In addition, for example, an identification method that focuses only on the time interval between characters in a voice uttered by a specific individual may be used. In this case, the system has a relatively low security level and can be realized at a low price.
Although the present invention has been described with reference to the above embodiment, the present invention is not limited to the above embodiment, and can be improved or changed within the scope of the purpose of the improvement or the idea of the present invention.

It is a figure which shows the structure of RF-ID card | curd and information processing apparatus which concern on embodiment of this invention. It is a figure which shows the structure of the data extraction storage device which concerns on embodiment of this invention. It is a flowchart which shows the flow of an individual identification which concerns on embodiment of this invention. It is a block diagram which shows the structure of the audio | voice data extraction means used for the collation regarding the audio | voice frequency which concerns on embodiment of this invention, and a collation means. It is a table | surface which illustrates the average value of the frequency peak to the 3rd of the vowel in the audio | voice frequency data which concerns on embodiment of this invention.

Explanation of symbols

10 RF-ID card (storage device)
11 Personal information storage means 12 Unique information storage means 20 Extraction storage device 40 Information processing device 43 Personal information / unique information confirmation means (reading means)
44 Microphone (voice input means)
45 voice data extracting means 46 collating means 50 door lock

Claims (10)

A portable storage device that stores a first time interval between a plurality of predetermined characters and a plurality of the characters in advance from a voice uttered by a specific individual for a predetermined series of characters;
Reading means for reading the first time interval and the plurality of characters stored in the storage device, voice input means for inputting voice uttered by the identified person, and the plurality of characters of the voice input by the voice input means Voice data extracting means for extracting a second time interval corresponding to the first time interval, and the second time interval extracted by the voice data extracting means and the first time interval read by the reading means An information processing apparatus having collation means for collating
A personal identification system comprising: A portable storage device that stores in advance a first time interval between a plurality of predetermined characters from a voice uttered by a specific individual for a predetermined series of characters, the plurality of characters, and a predetermined first frequency;
Reading means for reading the first time interval, the plurality of characters, and the first frequency stored in the storage device, voice input means for inputting a voice uttered by the identified person, and input by the voice input means Voice data extracting means for extracting a second time interval corresponding to the first time interval and a second frequency corresponding to the first frequency from the plurality of characters of the voice, and extracted by the voice data extracting means An information processing apparatus comprising: collation means for collating the second time interval and the second frequency with the first time interval and the first frequency read by the reading means;
A personal identification system comprising: 3. The personal identification system according to claim 1, wherein the first time interval between the predetermined plurality of characters is a time interval of speech of the first and last characters in the predetermined series of characters. 4. The personal identification system according to claim 2, wherein the first frequency is a highest frequency among predetermined formant frequencies included in the voice uttered by a specific individual for the predetermined series of characters. The specific individual utterance of the predetermined series of characters is performed a plurality of times, and the average value and standard deviation of the first time interval or the first frequency in each utterance are stored in the storage device. The personal identification system according to claim 4. The personal identification system according to claim 1, wherein the storage device is an RF-ID card, an RF-ID tag, or a contact IC card. The personal identification system according to claim 1, wherein the storage device and the information processing device have means for exchanging information wirelessly. The personal identification system according to claim 1, wherein the storage device includes an extraction storage device that stores the first time interval and the first frequency. 9. A storage device used in the personal identification system according to claim 1, further comprising means for generating drive power by electromagnetic induction. Storing a predetermined first time interval, the plurality of characters, and a predetermined first frequency in a portable storage device from a voice uttered by a specific individual for a predetermined series of characters;
Reading the first time interval, the plurality of characters, and the first frequency stored in the storage device;
Inputting voice uttered by the identified person;
Extracting a second time interval corresponding to the first time interval and a second frequency corresponding to the first frequency from the plurality of characters of the voice input by the voice input means;
Collating the second time interval and the second frequency extracted by the voice data extracting means with the first time interval and the first frequency read by the reading means;
A personal identification method comprising: