JP4037009B2 - Iris recognition device and identification system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a personal identification device using biological features, and more particularly to an iris recognition device using human iris.
[0002]
[Prior art]
Conventionally, an iris recognition device using an iris as a human biological feature has been used. This iris recognition apparatus registers user's iris data as iris dictionary data in advance and, when collating, compares the user's iris data acquired by the camera with the iris dictionary data to perform iris recognition. Is.
[0003]
Such an iris recognition device has two operation modes, a registration mode and a verification mode. The registration mode is a mode used when a user who wants to use the apparatus registers user iris data in the apparatus prior to using the apparatus.
[0004]
The collation mode is an operation mode when the apparatus is used in order for a registered user to receive a service using the apparatus.
[0005]
In iris recognition, 1: 1 verification in which the user inputs his / her ID code or the like prior to verification and verification is performed, and 1: n verification in which the user does not input an ID code or the like. However, for the sake of simplification, a 1: 1 verification will be described as an example.
[0006]
[Registration mode]
First, the user's ID code is input by an input means such as a keyboard or a card reader. Next, an image of the iris to be personally identified is captured. Then, the iris of one eye of the user is cut out. Here, the boundary between the outer edge and the inner edge of the iris is detected. Next, a predetermined feature amount is extracted from the cut image of the iris portion and coded. The encoded iris feature quantity of one eye of the user is linked with the user ID code as personal identification data and stored as iris dictionary data in the database.
[0007]
[Verification mode]
First, the user's ID code is input by an input means such as a keyboard or a card reader. Next, an image of the iris to be personally identified is captured. Then, the iris of one eye of the user is cut out. Here, the boundary between the outer edge and the inner edge of the iris is detected. Next, a predetermined feature amount is extracted from the extracted image of the iris portion and coded. The encoded iris feature value of one eye of the user is matched with matching iris data in the database. Specifically, the dissimilarity calculation is performed between the input feature data of the iris and the corresponding data in the registration data.
[0008]
The degree of dissimilarity obtained by the collation is determined using a certain threshold value to determine whether or not the person is the person. When the dissimilarity d obtained by the collation is equal to or smaller than a threshold D (d <D), the user is confirmed to be a registered person. On the other hand, if the degree of dissimilarity obtained by collation exceeds the threshold (d ≧ D), it is determined that the user is not the registered person and is rejected.
[0009]
[Problems to be solved by the invention]
However, in the conventional iris recognition apparatus, there is no particular problem in the case of a naked eye person during operation, but in the case of a person using glasses or a contact lens (particularly a hard contact lens), reflection of light by the glasses. In some cases, the acquired image of the eye may be deteriorated due to contact lens displacement or the like. In such a case, the dissimilarity slightly exceeds the threshold value and is rejected even though it is the person. There was a problem that rejection occurred and the reliability of processing as a device was lowered.
[0010]
[Means for Solving the Problems]
The present invention employs the following configuration in order to solve the above-described problems.
<Configuration 1>
The iris dictionary that stores the registered iris data of both eyes of the registrant, the iris data obtained from one of the user's eyes and the registered iris data of the iris dictionary are checked against the user, either the user or the other person A first determination unit that determines that neither the principal nor the other person can be determined, and the iris data acquired from the other eye of the user and the registered iris data of the iris dictionary when determined to be undecided An iris recognition apparatus, comprising: a second determination unit that collates and determines that the collation result is undecided.
[0011]
<Configuration 2>
A biological information registration unit biometric information sites having different subscribers are registered, and the acquired biometric information and the registered biometric information of the biometric information registration unit from one site of the user by matching the person to the user A first determination unit that determines that one of the other person and neither the person nor the other person can be determined, and biometric information acquired from the other part of the user and biometric information registration that is determined to be undecided A personal identification system comprising: a second determination unit that collates the registered biometric information of each unit and determines that the verification result is undecided.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail using specific examples.
<< Specific Example 1 >>
<Constitution>
FIG. 1 is a block diagram showing a specific example of the iris recognition apparatus of the present invention.
The apparatus shown in the figure includes a control unit 1, an optical unit 2, a display unit 3, an input unit 4, and an iris dictionary 5.
[0014]
The control unit 1 is a functional unit that performs overall control of the apparatus, image capture, and overall recognition processing, and includes a registration unit 11, a first determination unit 12, and a second determination unit 13. The controller 1 is specifically composed of a personal computer or the like. The registration unit 11 has a function of acquiring iris data of both eyes and registering the iris data of both eyes in the iris dictionary 5. The first determination unit 12 has a function of collating the iris data of any one eye of the user with the iris dictionary data in the iris dictionary 5 and sending the result of the identity determination result. If the result of the first determination unit 12 is negative, the second determination unit 13 performs the iris verification of the other eye, and if the result of the other eye verification is acceptable, the second determination unit 13 uses the result. The person has a function to determine that the person himself is correct.
[0015]
The optical unit 2 is a unit that captures an image of an iris to be recognized, and specifically includes an area sensor, a lens, illumination, and the like. The display unit 3 is a display means for displaying a recognition result, displaying a captured iris image, and the like, and specifically includes a CRT, a liquid crystal display, a status display LED, and the like. The input unit 4 is a display unit for inputting an ID code or the like to the apparatus, and specifically includes a keyboard and a card reader.
[0016]
The iris dictionary 5 is a database for storing the iris data of both eyes in association with the user's ID code, and specifically comprises a hard disk device or a semiconductor memory.
[0017]
<Operation>
Next, the operation of the iris recognition apparatus having the above configuration will be described. In this specific example, the 1: 1 recognition will be described as an example in iris recognition.
[0018]
[Registration mode]
FIG. 2 is an operation flowchart of the registration mode of the iris recognition apparatus according to the first specific example.
First, the user's ID code is input by an input means such as a keyboard or a card reader in the input unit 4 (step S1). Next, an image of the iris to be personally identified is captured by the optical unit 2 (step S2). Then, the iris of both eyes of the user is cut out. Here, the boundary between the outer edge and the inner edge of the iris is detected (step S3).
[0019]
A predetermined feature amount is extracted from the cut image of the iris portion and coded (step S4). The encoded iris feature quantity of both eyes of the user is linked with the ID code of the user as personal identification data by the registration unit 11 and stored as iris dictionary data in the database (step S5).
[0020]
Thus, in this specific example, the iris data of both eyes of the user is stored in the iris dictionary 5 in advance.
[0021]
[Verification mode]
In this specific example, when the collation is rejected by collation with one eye, the other eye is also collated, and the collation results of both eyes based on a plurality of threshold values (two threshold values in this specific example) are combined. Judgment is made to reduce the occurrence of rejection.
[0022]
FIG. 3 is an operation flowchart in the collation mode.
FIG. 4 is an explanatory diagram of the determination result of this example.
[0023]
In FIG. 4, Dl is the first threshold value, Dh of dissimilarity If you are sure that a person represents a second threshold of dissimilarity when other people almost certain, Dl <Dh relationship It has become. In the figure, x, Δ, and o indicate the following relationship.
[0024]
×: Dh <d (judgment result is another person)
Δ: Dl ≦ d ≦ Dh (judgment result is undecided)
○: d <Dl (judgment result is the person)
[0025]
First, the user's ID code is input by an input means such as a keyboard or a card reader in the input unit 4 (step S11). Next, an image of the iris to be personally identified is captured by the optical unit 2 (step S12).
[0026]
Next, the iris of one eye of the user (for example, the left eye in this specific example) is cut out. Here, the boundary between the outer edge and the inner edge of the iris is detected (step S13). Then, a predetermined feature amount is extracted from the cut image of the iris portion and coded (step S14).
[0027]
Next, the encoded user's iris feature amount is collated (matched) with the iris dictionary data corresponding to the ID code in the iris dictionary 5 by the first determination unit 12, and the degree of dissimilarity thereof is determined. Is calculated (step S15). In the collation process in step S15, if the left eye dissimilarity dl <Dl (indicated by ◯ in FIG. 4), it is determined that the person is himself (steps S16 and S17).
[0028]
On the other hand, if the left eye dissimilarity dl ≧ Dl (indicated by Δ or x in FIG. 4) in the determination process of step S16, the determination result in the first determination unit 12 is determined to be negative, and the other The operation proceeds to the matching operation for the first eye (in this case, the right eye).
[0029]
The iris of the user's other eye (right eye) is cut out. Here, the boundary between the outer edge and the inner edge of the iris is detected (step S18). Then, a predetermined feature amount is extracted from the cut out image of the iris portion and coded (step S19).
[0030]
Next, the encoded user's iris feature amount is collated (matched) with the iris dictionary data corresponding to the ID code in the iris dictionary 5 by the second determination unit 13, and the degree of dissimilarity thereof is determined. Is calculated (step S20). In the collation process of step S20, if the dissimilarity dr <Dl of the right eye (indicated by a circle in FIG. 4), the comprehensive determination result is the person (steps S21 and S22).
[0031]
Also, both eyes of humans have completely different iris patterns, and when both eyes have relatively low dissimilarity, it can be confirmed with the person himself, so both dissimilarities (dl, dr) of both eyes are Dl ≦ ( Also when dl, dr) ≦ Dh (indicated by Δ in FIG. 4), the comprehensive determination result is determined to be the person himself (steps S21 and S22). If it is a value other than this, it is determined that the user is not the person himself and is rejected (step S23).
[0032]
<effect>
As described above, according to the first specific example, the iris dictionary data of both eyes are prepared, and when the collation is rejected by the collation with one eye, the other eye is also collated, and a plurality of threshold values are obtained. The generation of rejection can be reduced by comprehensively judging the collation result of both eyes. As a result, for example, even if a person is wearing glasses or wearing a contact lens, the probability that the person is rejected can be reduced, and the reliability of the apparatus can be improved.
[0033]
<< Specific Example 2 >>
In specific example 2, when both eyes are collated, the user's iris data is stored in the database as iris dictionary data.
[0034]
<Constitution>
Since the configuration on the drawing is the same as the configuration of the specific example 1 shown in FIG. 1, the description will be made with reference to FIG.
This specific example is different from the specific example 1 in that it relates to the function of the registration unit 11. That is, the registration unit 11 in the specific example 2 has a function of registering the iris data of the user in the iris dictionary 5 as the iris dictionary data when the second determination unit 13 determines that the user is the person. Yes. Since each other structure is the same as that of the specific example 1, the description is abbreviate | omitted.
[0035]
<Operation>
[Registration mode]
Also in the second specific example, the normal registration mode is the same as that in the first specific example, and the description thereof is omitted.
[0036]
[Verification mode]
FIG. 5 is an operation flowchart of the collation mode of the second specific example.
The operation of the specific example 2 is different from the operation of the specific example 1 shown in FIG. 3 only in step S44, and steps S31 to S43 are the same as steps S11 to 23 in FIG.
[0037]
In specific example 2, when it is confirmed that the user is the person in step S42, the registration unit 11 determines that the iris data is determined not to be the person in step S36 (data of Δ and x in FIG. 4), or When the dissimilarity (dl, dr) of both eyes is Dl ≦ (dl, dr) ≦ Dh (data of Δ for both eyes in FIG. 4), the iris data is used as iris dictionary data. 5 (step S44).
[0038]
By performing such an operation, iris data having a certain degree of difference in dissimilarity is registered as iris dictionary data (so-called multi-template), so that the recognition rate can be improved in the next collation. For example, in the previous time, even if the left eye is × and the right eye is ○, this time it is compared with the iris data when it becomes × this time, so the probability that it can be confirmed by the first left eye verification can be increased. it can.
[0039]
<effect>
As described above, according to the second specific example, in addition to the configuration of the first specific example, when it is determined that the person is the result of the collation of both eyes, the iris data in which the person determination result is negative is determined as the iris data. Since it is registered as dictionary data, the recognition rate at the next collation can be improved, and as a result, the reliability of the iris recognition device can be improved.
[0040]
In each of the above specific examples, the identity determination is performed by comparing the dissimilarity with the first and second thresholds. However, the identity determination may be performed by comparing the similarity with a predetermined threshold. Good. Further, in each specific example, the 1: 1 collation has been described as an example, but the 1: n collation is similarly applicable.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a specific example 1 of an iris recognition apparatus of the present invention.
FIG. 2 is an operation flowchart of a registration mode of the iris recognition apparatus of specific example 1;
FIG. 3 is an operation flowchart of a collation mode of the iris recognition apparatus of specific example 1;
FIG. 4 is an explanatory diagram of a determination result of the iris recognition apparatus of specific example 1;
FIG. 5 is an operation flowchart of a collation mode of the iris recognition apparatus of specific example 2;
[Explanation of symbols]
5 Iris dictionary 11 Registration unit 12 First determination unit 13 Second determination unit

Claims (5)

An iris dictionary in which the registered iris data of both eyes of the registrant is stored;
It is undetermined that the iris data acquired from one eye of the user and the registered iris data in the iris dictionary are collated to determine whether the user is the person or the other person and neither the person nor the other person can be determined. a first determination unit determines that,
A second determination unit that compares the iris data acquired from the other eye of the user with the registered iris data of the iris dictionary when it is determined to be undecided, and determines that the user is undecided if the collation result is undecided; ,
An iris recognition device comprising:   2. The iris recognition according to claim 1, wherein the second determination unit determines that the verification result is the person himself / herself even if the first determination unit determines whether the person is undetermined or another person. apparatus.   2. The iris according to claim 1, further comprising: a registration unit that additionally registers two iris data acquired from the two eyes in the iris dictionary when the first and second determination units determine that the two are not yet determined. Recognition device.   The first determination unit and the second determination unit refer to the first threshold for determining the person and the second threshold for determining the other person, the person, the undecided, and the other person's The iris recognition device according to claim 1, wherein any one of them is determined. A biometric information registration unit in which biometric information of different parts of the registrant is registered;
The biometric information acquired from one part of the user and the biometric information registered in the biometric information registration unit are collated to determine whether the user is the person or the other person, and neither the person nor the other person is determined. And a first determination unit that determines to be undecided ,
A second determination unit that, when determined to be undecided, collates the biometric information acquired from the other part of the user with the registered biometric information of the biometric information registration unit and determines that the collation result is undecided When,
A personal authentication system comprising:

Images