JP6214908B2 - Verification device and verification method - Google Patents

Description

  The present invention relates to a collation apparatus and a collation method for collating registered data registered in advance with inputted collation data.

  In recent years, in areas where personal authentication is required, such as entry / exit management for computer rooms and critical machine rooms, and access control for computer terminals and financial terminals, fingerprint authentication and face cards have been used instead of personal identification numbers and ID cards. Biometric authentication such as authentication is being adopted. In a collation apparatus employing such biometrics authentication, face and fingerprint image data (hereinafter referred to as “registration data”) registered in advance and a face or fingerprint image input from a camera or sensor, etc. Identity verification is performed by collating data (hereinafter referred to as “collation data”) (see, for example, Patent Documents 1 and 2).

  In such a collation device, the registered data and the collation data are compared, the degree of similarity (hereinafter referred to as “similarity”) is quantified, the similarity is compared with a threshold value, and the similarity is compared. If the value exceeds the threshold, the person is certified.

JP 2008-066552 A JP 2013-003957 A

J. Ross Beverudge, David Bolme, Bruce A. Draper, Marcio Teixeira, The CSU Face Identification Evaluation System, Machine Vision and Applications (2005) 16: 128-138

  However, in the above-described method, if the similarity between the collation data of a person who is not registered in advance and any registration data exceeds a threshold value, the person is mistakenly recognized as a registered person even if it is not registered. In some cases, it continued to be misunderstood. In order to improve such a situation, if the threshold is tightened and the ratio of accepting others is kept low, the ratio of rejecting the person increases and the convenience of the user is sacrificed.

  Accordingly, an object of the present invention is to provide a collation device and a collation method that can prevent the recurrence of misidentification.

In order to solve the above-described problems, a collation device according to the present invention includes an input unit to which collation data is input, a first storage unit that stores a plurality of registration data in advance, and collation data that has been misidentified in the past. When the second storage unit stored in advance, the collation data and the registration data stored in the first storage unit are collated, and the registration data that matches the collation data is stored in the first storage unit, The first verification unit that determines that the verification data matches the registered data, the verification data that is determined to match by the first verification unit, and the verification data that is stored in the second storage unit If the matching data determined to match by the first matching unit is stored in the second storage unit, the registered data determined to match by the first matching unit does not match the matching data. Determine the first A second collation that determines that the registration data determined to match by the first verification unit and the verification data match if the verification data determined to match by the fixed unit is not stored in the second storage unit , A third storage unit that stores the registration data determined to match by the second verification unit and the verification data in association with each other, and the registration data stored in association with each other in the third storage unit And the collation data reach a predetermined quantity greater than 1, it is confirmed whether or not the registered data and collation data stored in association with each other in the third storage unit are the same person. And a confirmation unit that stores the collation data confirmed to be absent in the second storage unit .

In the verification device, the first verification unit compares the verification data input to the input unit with the registration data stored in the first storage unit, and the degree of similarity between the verification data and the registration data And a determination unit that determines that the registered data matches the collation data having the maximum similarity among the similarities exceeding a predetermined threshold value. May be.
Here, the calculation unit may calculate the similarity by a phase-only correlation method.

The collation method according to the present invention includes an input step for inputting collation data, a first storage step for storing a plurality of registration data in the first storage unit in advance, and a collation that has been misidentified in the past in the second storage unit. The second storage step for storing data in advance, the verification data and the registration data stored in the first storage unit are verified, and the registration data that matches the verification data is stored in the first storage unit. In this case, the first collation step for determining that the collation data and the registered data match, the collation data determined to match by the first collation step, and the collation data stored in the second storage unit If the matching data determined to match in the first matching step is stored in the second storage unit, the registered data determined to match in the first matching step Registration data determined to match with the first matching step when the matching data determined to match with the matching data and not matched with the first matching step is not stored in the second storage unit And a second storage step for storing in the third storage unit the registered data determined to match in the second determination step and the verification data in association with each other. When the registered data and the collation data stored in association with each other in the third storage unit reach a predetermined quantity greater than 1, the registration data and collation stored in the third storage unit in association with each other data and confirms whether or not the same person, and characterized by having a confirmation step of storing the verification data is not a same person has been confirmed in the second storage unit Is shall.

  According to the present invention, even if the collation data is determined to match, it is determined that the collation data does not match if the collation data misidentified in the past is included in the second storage unit that has been stored in advance. Therefore, the recurrence of misidentification can be prevented.

FIG. 1 is a block diagram showing the configuration of the collation apparatus according to the first embodiment of the present invention. FIG. 2 is a diagram schematically showing the structure of data stored in the misidentified data storage unit. FIG. 3 is a flowchart showing a registration operation by the collation device according to the first embodiment of the present invention. FIG. 4 is a flowchart showing a collation operation by the collation apparatus according to the first embodiment of the present invention. FIG. 5 is a flowchart showing a confirmation operation by the collation device according to the first embodiment of the present invention. FIG. 6 is a block diagram showing a configuration of a collation apparatus according to the second embodiment of the present invention. FIG. 7 is a flowchart showing a collation operation by the collation apparatus according to the second embodiment of the present invention.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 is a block diagram showing a configuration of a collation system including a collation apparatus according to the first embodiment of the present invention. The verification system shown in FIG. 1 includes, for example, a verification apparatus 1 installed in a building management room, and one or a plurality of receptions installed at the entrance of the building or the entrance of a room in the building. Based on the communication line 3 that connects the device 2, the verification device 1, and the reception device 2, the camera 4 that is installed in the vicinity of the reception device 2, transmits the captured image data to the reception device 2, and an instruction from the reception device 2 And an electric lock 5 for locking and unlocking the entrance and the door of the room.
Such a collation system transmits image data of a user who wants to enter or enter the room, which is captured by the camera 4, to the collation device 1 through the reception device 2, and a person included in the image data is identified by the collation device 1. It is confirmed whether the person is permitted to enter the building or room in which the receiving device 2 is installed, or whether the person is a person. Based on the result confirmed by the checking device 1 by the receiving device 2, the electric lock 5 is unlocked and unlocked. Is what you do.

<Configuration of collation device 1>
The collation device 1 includes a registration unit 11, a first collation unit 12, a second collation unit 13, a confirmation unit 14, and a storage unit 15.

<< Configuration of Registration Unit 11 >>
The registration unit 11 is a functional unit that performs a registration operation of registering a new user who permits entry into a building or room in which the reception device 2 is installed or a room. Such a registration unit 11 includes a registration data input unit 111.

  The registration data input unit 111 receives biometric authentication image data (registration data) of a newly registered person imaged by the camera 4 via the accepting device 2 and the communication line 3, and this registration data will be described later. The data is stored in the registration data storage unit 151 of the storage unit 15.

<< Configuration of First Collation Unit 12 >>
The first verification unit 12 includes image data (verification data) for biometric authentication of a user who desires to enter or enter a building or room in which the accepting device 2 is installed, and a registration data storage unit of the storage unit 15 described later. This is a functional unit that performs a collation operation for collating the registration data stored in 151 and confirming whether or not registration data that matches the collation data is stored in the registration data storage unit 151. Such a first collation unit 12 includes a collation data input unit 121, a collation data storage unit 122, a first calculation unit 123, a first determination unit 124, and a first threshold storage unit 125. And.

  The verification data input unit 121 receives image data (verification data) for biometric authentication of a person who wants to enter or enter the room, which is captured by the camera 4, via the receiving device 2 and the communication line 3. Hereinafter, the input data is also referred to as “input verification data”).

  The verification data storage unit 122 stores the verification data input from the verification data input unit 121.

  The first calculation unit 123 compares the collation data stored in the collation data storage unit 122 with each of the plurality of registration data stored in the registration data storage unit 151 of the storage unit 15, and the two are similar. The degree of similarity indicating the degree of the registration is calculated for each registered data. In the present embodiment, the first calculation unit 123 searches for the feature region of the matching data by the phase-only correlation method, compares the feature region with the corresponding region of the registered data, and the two are similar. A numerical value indicating the degree is calculated as the similarity. In the present embodiment, the similarity indicates that the higher the numerical value, the more similar.

The first determination unit 124 determines whether or not registration data that matches the verification data is included in the registration data storage unit 151 of the storage unit 15. In the present embodiment, the first determination unit 124 uses the first threshold value stored in the first threshold value storage unit 125 among the similarities calculated by the first calculation unit 123. The collation data having the maximum similarity among the exceeding degrees of similarity and the registration data are those of the same person, that is, registration data matching the collation data is included in the registration data storage unit 151 of the storage unit 15. Is determined. When it is determined that they match in this way, the first determination unit 124 outputs the verification data (input verification data) to the second verification unit 13.
On the other hand, when there is no similarity exceeding the first threshold in the similarity calculated by the first calculation unit 123, the first determination unit 124 receives the determination result via the communication line 3. Transmit to device 2.

  The threshold value storage unit 125 stores a first threshold value used in the first determination unit 124 in advance. This first threshold value is freely set as appropriate.

<< Configuration of Second Collation Unit 13 >>
The second collation unit 13 collates the collation data determined to be matched by the first collation unit 12 with the collation data included in the misidentification data storage unit 152 of the storage unit 15 described later, and the first collation This is a functional unit that performs a collation operation for confirming whether or not the collation data determined to match by the unit 12 is stored in the misidentification data storage unit 152. Such a second collation unit 13 includes a second calculation unit 131, a second determination unit 132, and a second threshold storage unit 133.

  The second calculation unit 131 misidentifies the collation data (input collation data) input from the first determination unit 124 of the first collation unit 12 and the past stored in the misidentification data storage unit 152 of the storage unit 15. The collation data (hereinafter also referred to as “misidentification collation data”) is compared, and a similarity indicating the degree of similarity between the two is calculated for each misidentification collation data. The calculation of the similarity can be performed by a method similar to the method for calculating the similarity by the first calculation unit 123 described above.

The second determination unit 132 determines whether or not the verification data (input verification data) determined to match by the first determination unit 124 of the first verification unit 12 is included in the misidentification data storage unit 152. . When the input verification data is not included in the misidentification data storage unit 152, the second determination unit 132 determines that the input verification data determined to match by the first determination unit 124 matches the registered data. On the other hand, when the input verification data is included in the misidentification data storage unit 152, the second determination unit 132 determines that the input verification data determined to match by the first determination unit 124 does not match the registered data. To do.
In the present embodiment, the second determination unit 132 exceeds the second threshold value stored in the second threshold value storage unit 133 among the similarities calculated by the second calculation unit 131. The input matching data having the maximum similarity matches the misidentification matching data, that is, the input matching data is included in the misidentification data storage unit 152, and therefore the input determined to be coincident by the first determination unit 124 It is determined that the verification data does not match the registered data.
On the other hand, if there is no input collation data and misidentification collation data having a similarity exceeding the second second threshold among the similarities calculated by the second calculation unit 131, the input collation data is stored as misidentification data. Since it is not included in the unit 152, it is determined that the input collation data determined to match by the first determination unit 124 matches the registered data.
The determination result by the second determination unit 132 is transmitted to the accepting device 2 via the communication line 3.
When it is determined that they match, the second determination unit 132 associates the input verification data and the registration data with each other and stores them in the match data storage unit 153 of the storage unit 15 described later.

  The second threshold value storage unit 133 stores a second threshold value used in the second determination unit 132 in advance. This second threshold value is freely set as appropriate.

<< Configuration of Confirmation Unit 14 >>
The confirmation unit 14 is a person registered in advance by the first collation unit 12, and whether or not the user who is determined not to be misidentified by the second collation unit 13 is a person registered in advance. It is a functional unit that performs a confirmation operation for confirming. Such a confirmation unit 14 includes a third determination unit 141.

The third determination unit 141 determines whether the collation data stored in association with the coincidence data storage unit 153 of the storage unit 15 to be described later and the registered data are the same person. Here, the third determination unit 141 makes a determination using a method different from the method determined by the first determination unit 124. In the present embodiment, if the third determination unit 141 performs collation using the phase-only correlation method in the first determination unit 124, for example, a conventional two-dimensional correlation method using amplitude information, Then, the feature point of the image is extracted, and the feature amount is calculated from the surrounding area, and the determination is performed by using the feature extraction method for matching between the images. Note that the third determination unit 141 may display the collation data and the registration data on the display device, and determine whether or not they are the same person by visual observation of the user.
If there is collation data and registration data that are not the same person, the third determination unit 141 stores the collation data in a misidentification data storage unit 152 of the storage unit 15 described later.

<< Configuration of Storage Unit 15 >>
The storage unit 15 is a functional unit that stores information necessary for various operations of the verification device 1. Such a storage unit 15 includes a registration data storage unit 151, a misidentification data storage unit 152, and a match data storage unit 153.

  The registration data storage unit 151 stores registration data input from the registration data input unit 111. This registration data includes biometric information of a person who is permitted to enter or enter the room.

  The misidentification data storage unit 152 stores collation data (misidentification collation data) input from the confirmation unit 14. This collation data is collation data that has been confirmed by the confirmation unit 14 to have been mistaken in the past. An example of the data configuration stored in the misidentification data storage unit 152 is shown in FIG. According to FIG. 2, the misidentification data storage unit 152 includes collation data (biometric information) including face image data. In addition, the misidentification data storage unit 152 includes the identification number of registration data misidentified as matching, the identification number of collation data, the time when collation data was input, the number of misidentifications, information on the time spent in a building or a room, and the like. Alternatively, registration may be performed in association with each verification data.

  The coincidence data storage unit 153 stores collation data and registration data input from the second determination unit 132 of the second collation unit 13 in association with each other. The collation data and the registration data coincide with each other by the first collation unit 12 and the second collation unit 13, that is, it is determined that the collation data and the registration data are the same person.

Such a collation apparatus 1 includes an arithmetic unit such as a CPU, a storage device such as a memory and an HDD (Hard Disc Drive), and an input that detects input of information from the outside such as a keyboard, mouse, pointing device, button, and touch panel. A device, an I / F device that transmits and receives various information via a communication line such as the Internet, a LAN (Local Area Network), and a WAN (Wide Area Network), and a display device such as an LCD (Liquid Crystal Display) It consists of a computer and a program installed on this computer. That is, the hardware device and software cooperate to control the hardware resources by a program, and the registration unit 11, the first verification unit 12, the second verification unit 13, the confirmation unit 14, and the storage described above. Unit 15 is realized.
Note that the program may be provided in a state of being recorded on a recording medium such as a CD-ROM, a DVD-ROM, or a memory card.

<Configuration of accepting device 2>
The receiving device 2 includes an image acquisition unit 31, a lock control unit 32, and a registration unit 33.

  The image acquisition unit 31 includes an interface device that acquires an image captured by the camera 4 attached in the vicinity of the entrance or door.

  The lock control unit 32 includes a control device that locks / unlocks the electric lock 5 provided at the entrance or the door based on the determination result received from the verification device 1.

  The registration unit 33 includes an interface device that performs a registration operation of registration data of a person who is newly permitted to enter or enter the room.

<Configuration of communication line 3>
The communication line 3 includes the Internet, an intranet, a public line, a dedicated line, or the like.

<Configuration of camera 4>
The camera 4 is composed of a digital camera that records an image captured by the image sensor as digital data.

<Configuration of electric lock 5>
The electric lock 5 includes a lock system that electrically locks and unlocks a lock provided on the door.

[Operation of verification device]
Next, operation | movement of the collation apparatus 1 is demonstrated. The operation of the verification device 1 is broadly divided into a registration operation for registering registration data by the registration unit 11, a verification operation by the first verification unit 12 and the second verification unit 13, and a verification operation by the verification unit 14. . Each will be described below.

<Registration operation>
First, the registration operation by the registration unit 11 will be described with reference to FIG.

  A new user who is permitted to enter or enter a building or room where the reception device 2 is installed registers registration data via the reception device 2. Specifically, the input device of the accepting device 2 is operated, the camera 4 is imaged and necessary information is input. Then, the registration unit 33 of the reception device 2 sends the image data of the new user acquired by the image acquisition unit 31 from the camera 4 and a control signal for requesting registration using the image data as registration data. 1 to send. As a result, when registration data and a control signal are received from the accepting device 2 (step S1: YES), the registration data input unit 111 of the registration unit 11 uses the received image data as registration data as a registration data storage unit 151 of the storage unit 15. (Step S2). By repeating such a registration operation, a plurality of registration data are registered in the registration data storage unit 151.

<Verification operation>
Next, the collating operation by the first collating unit 12 and the second collating unit 13 will be described with reference to FIG.

  A user who wants to enter or enter a building or room in which the receiving apparatus 2 is installed in a state where a plurality of registration data is registered in the registration data storage unit 151 of the storage unit 15 in advance. To image. Then, the image acquisition unit 31 of the reception device 2 transmits image data captured by the camera 4 to the verification device 1. When image data is received from the accepting device 2 (step S11: YES), the collation data input unit 121 of the first collation unit 12 registers the image data in the collation data storage unit 122 as collation data.

  When the collation data is registered, the first calculation unit 123 of the first collation unit 12 uses the collation data stored in the collation data storage unit 122 and the registration data stored in the registration data storage unit 151. Similarity is calculated by sequentially comparing (step S12).

  When the similarity is calculated, the first determination unit 124 of the first matching unit 12 stores the similarity calculated by the first calculation unit 123 in the first threshold value storage unit 125. It is confirmed whether or not there is collation data and registration data having a similarity exceeding the first threshold (step S13).

  When there is no collation data and registration data whose similarity exceeds the first threshold (step S13: NO), the first determination unit 124 does not have any registration data that matches the collation data (collation NG). (Step S18). This determination result is transmitted to the lock control unit 32 of the accepting device 2 by the first determination unit 124. When receiving the determination result “verification NG”, the lock control unit 32 does not unlock the electric lock 5. Thereby, it is possible to prevent a user whose registration data is not registered from entering or entering the room.

  On the other hand, when there is collation data and registration data whose similarity exceeds the first threshold (step S13: YES), the first determination unit 124 determines the similarity that exceeds the first threshold. It is determined that the matching data having the maximum similarity matches the registered data.

  When the first determination unit 124 determines that the collation data and the registered data match, the second calculation unit 131 of the second collation unit 13 matches the collation data (input collation data) determined to match. Then, the collation data (misidentification collation data) stored in the misidentification data storage unit 152 is compared to calculate the similarity for each misidentification collation data (step S14).

  When the similarity is calculated, the second determination unit 132 of the second collation unit 13 stores the similarity calculated by the second calculation unit 131 in the second threshold storage unit 133. It is checked whether or not there is input matching data having a similarity degree exceeding the second threshold and misidentification matching data (step S15).

  When there is input collation data and misidentification collation data whose similarity exceeds the second threshold value (step S15: YES), the second determination unit 132 is determined to match by the first determination unit 124. Since the input verification data is included in the misidentification data storage unit 152, that is, the person corresponding to the input verification data is the same person as the person misidentified in the past, there is no registration data that matches the input verification data (verification NG) (step S18). Thus, even if it is determined to collate with the registered data, if it is misidentified in the past, it is determined as collation NG, and thus the recurrence of misidentification can be prevented.

  On the other hand, when there is no input collation data and misidentification collation data whose similarity exceeds the second threshold value (step S15: NO), the second determination unit 132 agrees with the first determination unit 124. It is determined that the person corresponding to the determined input verification data is not a person mistaken in the past, and the registration data determined to be verified by the first determination unit 124 matches the verification data, that is, matches the verification data. It is determined that the registered data exists (check OK) (step S16). This determination result is transmitted to the lock control unit 32 of the accepting device 2 by the second determination unit 132. The lock control part 32 will unlock | release the electric lock 5, if the determination result called collation OK is received. Thereby, a user whose registration data is registered in advance can enter or enter the room.

  Further, the second determination unit 132 outputs the registration data determined to be verified and the verification data to the confirmation unit 14. The output registration data and collation data are registered in the coincidence data storage unit 153 in association with each other (step S17). As a result, the matching data storage unit 153 stores the registered data determined as collation OK and the collation data in association with each other.

<Confirmation action>
Next, the confirmation operation by the confirmation unit 14 will be described with reference to FIG.

  For example, if the security officer reports that a suspicious person has entered the room, the verification operation described above is performed only for a predetermined period, or if a predetermined amount of registration data and verification data are stored in the match data storage unit 153, The third determination unit 141 of the unit 14 checks whether or not the registration data and the collation data stored in association with the match data storage unit 153 are the same person (step S21).

  When registration data and verification data that are not the same person are stored in association with each other in the match data storage unit 153 (step S22: YES), the third determination unit 141 stores the verification data in the storage unit 15 Is stored in the misidentification data storage unit 152 (step S23). As a result, in the misidentification data storage unit 152, verification data that has been misidentified as not matching the registered data registered in the registration data storage unit 151 is registered. It can prevent misidentification.

  If registration data that is not the same person and matching data are not stored in association with each other in the coincidence data storage unit 153 (step S22: NO), the third determination unit 141 ends the misidentification data registration operation.

  At this time, the confirmation by the third determination unit 141 confirms the correctness of the determination result by the first determination unit 124, and thus is performed by a method different from that of the first determination unit 124. Therefore, even if the first determination unit 124 determines that the person is the same person, the third determination unit 141 may determine that the person is not the same person. Thereby, recurrence of misidentification can be prevented.

  As described above, according to the present embodiment, when the verification data determined to match by the first determination unit 124 is included in the misidentification data storage unit 152, the second determination unit 132 Since it is determined that the matching data determined to match does not match, the recurrence of misidentification can be prevented.

  Further, the third determination unit 141 confirms whether or not the collation data determined to match by the first determination unit 124 and the second determination unit 132 and the registered data are the same person. When it is confirmed that there is no data, the verification data is stored in the misidentification data storage unit 152, so that it is possible to prevent the person corresponding to the verification data from being misidentified again.

  Further, the phase only correlation method used in the first calculation unit 123 is known to have high identification performance (see, for example, Non-Patent Document 1). For this reason, even if another person's misidentification occurs, if the misidentified data is registered as in the present embodiment, it will match with the misidentified data in subsequent verifications, thus preventing the recurrence of misidentification by others. it can.

  In the present embodiment, the case where facial image data is used as image data for biometric authentication has been described as an example. However, image data for biometric authentication is not limited to this, and for example, fingerprint, retina, vein Such image data may be used.

  In addition, the verification device 1 may not be installed in the same building as the reception device 2 as long as it is connected to the reception device 2 via the communication line 3.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. Note that the present embodiment is different from the first embodiment described above, except that the configuration of the first determination unit and the second determination unit of the verification device and the verification operation by them are different. It is equivalent. Therefore, configurations and operations equivalent to those of the first embodiment are denoted by the same names and symbols, and description thereof is omitted as appropriate.

  The collation system shown in FIG. 6 includes, for example, a collation apparatus 1 ′ installed in a building management room, and one or a plurality of receptions installed at the entrance of the building or the entrance of a room in the building. A communication line 3 that connects the device 2, the verification device 1 ′, and the reception device 2; a camera 4 that is installed near the reception device 2 and transmits captured image data to the reception device 2; and an instruction from the reception device 2. And an electric lock 5 for locking and unlocking the entrance and the door of the room.

<Configuration of verification device 1 '>
The collation device 1 ′ includes a registration unit 11, a first collation unit 12 ′, a second collation unit 13 ′, a confirmation unit 14, and a storage unit 15.

<< Configuration of First Collation Unit 12 '>>
The first verification unit 12 ′ stores biometric authentication image data (input verification data) of a user who desires to enter or enter a building or room where the receiving device 2 is installed, and registration data storage in the storage unit 15. Whether the registration data stored in the unit 151 and the misidentification collation data stored in the misidentification data storage unit 152 are collated, and the registration data or the misidentification collation data matching the input collation data is stored in the storage unit 15. It is a functional unit that performs a collation operation for confirming. Such a first collation unit 12 includes a collation data input unit 121, a collation data storage unit 122, a first calculation unit 126, a first determination unit 127, and a first threshold storage unit 125. And.

  Here, the first calculation unit 126 is configured to input collation data stored in the collation data storage unit 122, registration data stored in the registration data storage unit 151 of the storage unit 15, and misidentification data in the storage unit 15. The misidentification collation data stored in the storage unit 152 is compared with each other, and a similarity indicating the degree of similarity between the two is calculated for each registered data. In the present embodiment, the first calculation unit 126 searches for a feature region of input collation data by the phase-only correlation method, compares this feature region with a corresponding region of registered data or misidentification collation data, A numerical value indicating the degree of similarity is calculated as the similarity. In the present embodiment, the similarity indicates that the higher the numerical value, the more similar.

The first determination unit 127 determines whether registration data or misidentification collation data that matches the input collation data is included in the registration data storage unit 151 or the misidentification data storage unit 152 of the storage unit 15. In the present embodiment, the first determination unit 127 uses the first threshold value stored in the first threshold value storage unit 125 among the similarities calculated by the first calculation unit 123. The input collation data having the maximum similarity among the exceeding degrees of similarity and the registration data or the misidentification collation data belong to the same person, that is, the registration data matching the input collation data is the registration data storage unit of the storage unit 15 151 or the misidentification data storage unit 152. When it determines with matching in this way, the 1st determination part 127 outputs collation data (input collation data) to 2nd collation part 13 '.
On the other hand, when the similarity calculated by the first calculation unit 126 does not have a similarity exceeding the first threshold, the first determination unit 127 receives the determination result via the communication line 3. Transmit to device 2.

<< Configuration of Second Matching Unit 13 '>>
The second collation unit 13 ′ includes data determined by the first collation unit 12 to match the input collation data, data included in the registration data storage unit 151 and the misidentification data storage unit 152 of the storage unit 15, and Whether the input verification data matches the registered data by checking whether the data determined to match the input verification data by the first verification unit 12 is the registration data or the misidentification verification data. It is a functional unit that performs a collation operation for determining whether or not. Such a second collation unit 13 includes a second determination unit 134.

In the present embodiment, the second determination unit 134 includes the data determined to match the input verification data by the first verification unit 12 ′ and the registration data included in the misidentification data storage unit 152 of the storage unit 15. And the first collation unit 12 ′ determines whether the data determined to match the collation data is registered data.
When the data determined to match the input verification data is not included in the misidentification data storage unit 152, the second determination unit 134 registers the data determined to match the input verification data by the first determination unit 127. Judged as data. On the other hand, when the data determined to match the input verification data is included in the misidentification data storage unit 152, the second determination unit 134 determines the data determined to match the input verification data by the first determination unit 127. Is not registered data.
The determination result by the second determination unit 134 is transmitted to the accepting device 2 via the communication line 3.
When the data determined to match the input collation data is registered data, the second determination unit 134 associates the input collation data and the registration data with each other and matches data storage unit of the storage unit 15 described later. 153 to store.

In the present embodiment, the second determination unit 134 confirms whether or not the misidentification data storage unit 152 includes data determined by the first determination unit 127 to match the input collation data. However, the storage unit used for the confirmation is not limited to the registration data storage unit 151. For example, the misidentification data storage unit 152 or the registration data storage unit 151 and the misidentification data storage unit 152 may be confirmed.
<Verification operation>
Next, the collating operation by the first collating unit 12 ′ and the second collating unit 13 ′ will be described with reference to FIG.

  When image data is received from the accepting device 2 (step S11: YES) and the image data is registered in the collation data storage unit 122 as input collation data by the collation data input unit 121, the first calculation unit 126 performs collation. Similarity is calculated by sequentially comparing the input collation data stored in the data storage unit 122 with the registration data stored in the registration data storage unit 151 and the misidentification collation data stored in the misidentification data storage unit 152. (Step S31).

  When the similarity is calculated, the first determination unit 127 uses the first threshold value stored in the first threshold value storage unit 125 among the similarities calculated by the first calculation unit 123. It is checked whether or not there is input collation data having a degree of similarity exceeding, registration data, or misidentification collation data (step S13).

  When there is no input collation data and registration data or misidentification collation data whose similarity exceeds the first threshold (step S13: NO), the first determination unit 127 registers data that matches the input collation data. Alternatively, it is determined that there is no misidentification collation data (collation NG) (step S18). This determination result is transmitted to the lock control unit 32 of the receiving device 2 by the first determination unit 127. When receiving the determination result “verification NG”, the lock control unit 32 does not unlock the electric lock 5. Thereby, it is possible to prevent a user whose registration data is not registered from entering or entering the room.

  On the other hand, when there is input collation data whose registration degree exceeds the first threshold value and registered data or misidentification collation data (step S13: YES), the first determination unit 127 uses the first threshold value. It is determined that the matching data having the highest similarity among the similarities exceeding the threshold value matches the registered data or the misidentified matching data.

  When the first determination unit 127 determines that the verification data matches the registered data or the misidentification verification data, the second determination unit 134 of the second verification unit 13 ′ determines that the input verification data matches. Whether the data determined by the first determination unit 127 to match the verification data is the registration data. It is determined whether or not (step S32).

  When data determined to match the input verification data is included in the misidentification data storage unit 152, the second determination unit 134 registers the data determined to match the input verification data by the first determination unit 127. It is determined that it is not data (step S32: NO). In this case, the second determination unit 134 is the same as the data that the first determination unit 124 determines to match the input collation data as the misidentification data, that is, the person that the person corresponding to the input collation data has previously misidentified. Since it is a person, it is determined that there is no registered data that matches the input verification data (verification NG) (step S18). Thus, since it is determined as collation NG when it has been misidentified in the past, it is possible to prevent the recurrence of misidentification.

  On the other hand, when the data determined to match the input verification data is not included in the misidentification data storage unit 152, the second determination unit 134 determines the data determined to match the input verification data by the first determination unit 127. Is registered data (step S32: YES). In this case, since the person corresponding to the data determined to match the input collation data by the first determination unit 124 is not a person mistaken in the past, the second determination unit 134 does not have the registration data that matches the collation data. It is determined that it exists (checking OK) (step S16). This determination result is transmitted to the lock control unit 32 of the accepting device 2 by the second determination unit 132. The lock control part 32 will unlock | release the electric lock 5, if the determination result called collation OK is received. Thereby, a user whose registration data is registered in advance can enter or enter the room.

  In addition, the second determination unit 134 outputs the registration data determined as verification OK and the verification data to the confirmation unit 14. The output registration data and collation data are registered in the coincidence data storage unit 153 in association with each other (step S17). As a result, the matching data storage unit 153 stores the registered data determined as collation OK and the collation data in association with each other.

  As described above, according to the present embodiment, even if the data is determined to match the input collation data, this data is not registered data, that is, misidentification in which collation data previously misidentified is stored in advance. When it is included in the data storage unit 152, it is determined that they do not match, so that the recurrence of misidentification can be prevented.

  In the present embodiment, the calculation of the similarity can be performed only by the first matching unit 12 ′, so that the calculation load can be reduced.

  In the present embodiment, the first determination unit 127 determines that the input collation data having the maximum similarity among the similarities exceeding the first threshold value matches the registered data or the misidentification collation data. Although the case of determination has been described as an example, it may be determined that the input collation data whose similarity exceeds the first threshold value and the registration data or the misidentification collation data all match. In this case, the second determination unit 134 determines that the verification is OK if the registered data is included in the data. Thereby, registration data that matches the input verification data can be detected.

  The present invention can be applied to various verification devices that perform identity verification.

  DESCRIPTION OF SYMBOLS 1,1 '... Collation apparatus, 2 ... Reception apparatus, 3 ... Communication line, 4 ... Camera, 5 ... Electric lock, 11 ... Registration part, 12, 12' ... 1st determination part, 13, 13 '... 2nd 14 ... confirmation part, 15 ... storage part, 31 ... image acquisition part, 32 ... lock control part, 33 ... registration part, 111 ... registration data input part, 121 ... collation data input part, 122 ... collation data storage Part, 123, 126 ... first calculation part, 124, 127 ... first determination part, 125 ... first threshold value storage part, 131 ... second calculation part, 132, 134 ... second determination part 133, second threshold value storage unit, 141, third determination unit, 151, registered data storage unit, 152, misidentified data storage unit, 153, coincidence data storage unit.

Claims (4)

An input unit for input of collation data;
A first storage unit storing a plurality of registration data in advance;
A second storage unit that previously stores collation data misidentified in the past;
The collation data and the registration data stored in the first storage unit are collated, and when the registration data matching the collation data is stored in the first storage unit, the collation data and A first verification unit that determines that the registration data matches,
The verification data determined to match by the first verification unit and the verification data stored in the second storage unit are verified, and the verification data determined to match by the first verification unit Is stored in the second storage unit, the registration data determined to be matched by the first matching unit is determined not to match the matching data, and is matched by the first determining unit. Then, when the collation data determined is not stored in the second storage unit, the second data is determined to be coincident with the registration data determined to be coincident with the collation data by the first collation unit. A collation unit ;
A third storage unit that stores the registered data and the verification data that are determined to be matched by the second verification unit in association with each other;
When the registration data and the collation data stored in association with each other in the third storage unit reach a predetermined quantity greater than 1, the registration data stored in association with each other in the third storage unit And a verification unit that checks whether or not the verification data is the same person, and stores the verification data that is verified to be not the same person in the second storage unit. Verification device. The verification device according to claim 1,
The first verification unit includes:
The registration data indicating the degree of similarity between the verification data and the registration data by comparing the verification data input to the input unit and the registration data stored in the first storage unit. A calculation unit to calculate for each,
A determination unit that determines that the collation data having the maximum similarity among the similarities exceeding a predetermined threshold matches the registered data;
Collating apparatus comprising: a. The verification device according to claim 2 ,
The calculation unit calculates the similarity by a phase-only correlation method.
A collation device characterized by that. An input step for entering verification data;
A first storage step of storing a plurality of registration data in the first storage unit in advance;
A second storage step of previously storing collation data misidentified in the past in the second storage unit;
The collation data and the registration data stored in the first storage unit are collated, and when the registration data matching the collation data is stored in the first storage unit, the collation data and A first matching step for determining that the registered data matches;
The collation data determined to match in the first collation step is collated with the collation data stored in the second storage unit, and the collation data determined to match in the first collation step Is stored in the second storage unit, it is determined that the registration data determined to match in the first verification step and the verification data do not match, and matches in the first verification step. Then, if the determined collation data is not stored in the second storage unit, a second determination is made that the registered data determined to coincide with the first collation step matches the collation data. A determination step ;
A third storage step of storing the registered data and the collation data determined to match in the second determination step in association with each other in a third storage unit;
When the registration data and the collation data stored in association with each other in the third storage unit reach a predetermined quantity greater than 1, the registration data stored in association with each other in the third storage unit And a step of confirming whether or not the collation data is the same person, and storing the collation data confirmed to be not the same person in the second storage unit. Method.

Images