JP2001043374A - Operation propriety decision device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent an incorrect decision of an operation due to the change of a photographing environment, etc., and accurately decide the propriety of an operation for an operation propriety decision device which compares the current image data obtained by photographing an operator with the registering image data which are previously recorded to decide the propriety of log-in to a personal computer. SOLUTION: In this operation propriety decision device, an area including the face part of an operator is defined as an area 82 to be seriously considered in an effective image area 80 of a photographed image and divided into small blocks 84... smaller than the blocks 85... of other areas. Then every block 84 is compared with the registering image data. Thus, the image information on the area 82 is relatively considered as a serious area and accordingly the points of difference of the face part of the operator are sensitively detected to improve the comparison accuracy. Meanwhile, other face parts of the operator are relatively despised and accordingly the minor points of difference such as a background part are granted to prevent the incorrect decision of propriety of an operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an operation permission / non-permission judging device for judging permission / non-permission of a personal computer (hereinafter, referred to as a personal computer) based on collation of current image data and registered image data based on imaging.

[0002]

2. Description of the Related Art In general, it is not preferable in terms of security and the like that a person other than a specific operator performs unlimited operations such as login on a personal computer. In accordance with this, permission or rejection of login or the like is performed.

[0003] Further, in a lock system such as an entrance of a research laboratory or a safe where confidentiality is required, it is determined whether or not a specific operator is determined by an IC card, a password, or the like. It is known to do so.

[0004]

Many conventional devices of this type make a determination using a password or the like. For example, an image of an operator's face is taken by a camera as an object to be determined, and the image data is obtained. The security is further improved by collating with the image data of a pre-registered valid operator, and determining whether or not the operator is a valid operator based on the collation, and determining whether or not to permit the operation accordingly. Can be expected.

[0005] However, normally, when the object to be determined is imaged, the background or the like is reflected. However, if the background or the like changes from the time when the image registered in advance is captured, the image data registered for the background portion and the image data registered for the background portion are not changed. Since the matching with the captured image data does not match, even if the object to be determined does not change, the matching matching rate of the entire image falls below a predetermined determination criterion, and the matching result is easily erroneously determined to be mismatch. There was a problem.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems, and it is possible to prevent an erroneous determination due to a change in an imaging environment or the like, and to perform an operation permission determination based on a highly accurate and reliable collation. It is intended to provide a device.

[0007]

According to the present invention, there is provided an operation permission / non-permission determining apparatus, comprising: an image pickup unit for picking up an object to be determined; and an image processing unit for performing image processing on an image picked up by the image pickup unit to create image data. Means, a registered dictionary in which the image data of the object to be determined is recorded in advance, and matching current image data captured by the image pickup unit and created by the image processing means with image data recorded in the registered dictionary. And a collation determining unit that determines whether the operation of the operation target is permitted based on the collation result, wherein the image processing unit is configured to divide the effective image area of the captured image into a plurality of blocks. After performing a division process of dividing the block, a block characteristic value representing each block is obtained for each block, and a set of the block characteristic values is created as image data. The division processing is characterized in that a predetermined area is specified as an important area from an effective image area of a captured image, and blocks in this important area are divided into smaller areas than blocks in other areas. .

According to such a configuration, of the effective image area of the image picked up by the image pickup unit, a predetermined area is specified as an important area, and the important area is divided into blocks of smaller area of other areas. Then, image data in which the image information in the specific area is more important than the image information in other areas is created. Therefore, if, for example, the area occupied by the object to be determined or the area occupied by the characteristic portion of the object to be determined is regarded as the important area, the difference between the current image data and the image data of the registered dictionary in these parts is sensitively detected. In addition, it is possible to realize more accurate and accurate collation, and to allow a slight difference between the two image data due to a change in an imaging environment such as a background portion, thereby preventing erroneous determination.

In such an operation permission / prohibition determining device,
The collation determination means may be configured to compare and collate block characteristic values of a plurality of blocks included in current image data and recorded image data for each corresponding block.

[0010] In this case, since the current image data and the image data in the registered dictionary are compared for each block, a highly accurate collation result can be obtained.

Further, in such an operation permission / prohibition judging device, before the collation by the collation judging means, the block characteristics of some of the blocks included in the current image data and the recorded image data are determined. It is preferable to perform a preliminary collation for comparing and collating the value for each corresponding block, and to perform the collation by the collation determination unit only when the preliminary collation result satisfies a predetermined preliminary collation determination standard. .

In this way, for the image data of the registered dictionary which is determined to be unsuitable by the preliminary collation for comparing and collating only some blocks, it is not necessary to compare and collate the current image data with all the blocks. The determination of non-conformity can be made in a short time.

Further, when a plurality of image data are recorded in the registration dictionary, image data to be subjected to the collation processing can be narrowed down by preliminary collation prior to the collation processing for comparing all blocks. As a whole, the time required for the collation determination processing can be reduced.

In such an operation permission / prohibition judging device, the image processing means performs a division process of dividing an effective image area of a captured image into a plurality of blocks, and then represents each block for each block. A block characteristic value is obtained, and further, a representative value representing the entire captured image is calculated as image data from a set of block characteristic values of each block, and the collation determination means is recorded as current image data. It is also possible to configure so as to compare and match a representative value with image data.

With this configuration, the captured image can be represented by one representative value while emphasizing the predetermined important area, so that the collation processing between the current image data and the image data recorded in the registration dictionary is extremely performed. It can be done easily.

In the division processing, it is preferable that the position of the important area is movable.

With this configuration, according to the position of the object to be determined in the captured image, the position of the characteristic portion of the object to be determined, and the like,
Since a region to be emphasized in the image can be specified as the important region, more accurate and accurate collation can be realized, and erroneous determination due to a change in the background portion can be prevented.

In the division processing, it is desirable that the size of the important area can be enlarged or reduced.

With this configuration, an area to be emphasized in the image is specified as the important area according to the size of the object to be determined in the captured image, the size of the characteristic portion of the object to be determined, and the like. Therefore, it is possible to realize more accurate and accurate collation, and to prevent erroneous determination due to a change in the background portion or the like.

In the division processing, it is possible to perform a deformation operation in which the important region is a region that is rotated by a predetermined angle around at least one axis with respect to the captured image plane and is a mapped area on the captured image plane. Is desirable.

With this configuration, the shape of the region to be emphasized in the image can be adjusted according to the orientation of the object to be determined in the captured image, the orientation of the characteristic portion of the object to be determined, and the like. In addition, it is possible to realize highly accurate and accurate collation, and to prevent erroneous determination due to a change in the background portion or the like.

The movement of the position of the important area, the change of the size, the deformation of the shape, and the like can be automatically performed by the information processing function of the operation permission / prohibition determination device, or can be changed by an operator's instruction. It may be performed based on.

Further, in such an operation permission / prohibition judging device, there is provided a monitor for displaying an image picked up by the image pickup section, and the important area is displayed in the image picked up on the monitor. Is desirable.

With this configuration, the operator can check the position, size, shape, and the like of the important area in the captured image by looking at the monitor. Therefore, according to the important area displayed on the monitor, By changing the position of the object to be determined with respect to the camera, or by moving the position of the important area in the captured image, the area to be emphasized in the image is specified as the important area, and more accurate and accurate matching is performed. This can be realized, and erroneous determination due to a change in the background portion can be prevented.

The display mode of the important area includes a frame line surrounding the important area, marks indicating four corners of the important area, and the like.
Alternatively, various forms can be adopted as long as the operator can confirm the important region in the captured image displayed on the monitor, such as a mark indicating the center position of the important region.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIGS. 1 and 2 show an overall configuration of an embodiment in which an operation permission / non-permission judgment apparatus according to the present invention is applied to a personal computer log-in permission / non-permission judgment apparatus. It is a block diagram of. In these figures, reference numeral 1 denotes a camera device which is a main body of the determination device, 5 denotes a personal computer to be operated, and 6 denotes a communication means for connecting the camera device 1 and the personal computer 5.

The communication means 6 comprises an interface 61 provided on the camera device 1 side, an interface 62 provided on the personal computer side, and a communication cable 60 for connecting these. However, the communication means 6 may be wireless using infrared rays or the like instead of using such a cable.

The camera device (determination device main body) 1 is connected to the signal control unit 10, the imaging unit 20, the image calculation unit 30, the working memory unit 41, the data holding memory unit 42, the program memory unit 11, and the personal computer 5. An interface 61 and the like are provided.

The signal control section 10 forms a central control section for controlling a series of operations in the camera device 1 which is the main body of the determination device. For this reason, the signal control unit 10 is connected to each unit of the camera device 1 and the program memory unit 11
These operations are controlled in accordance with the procedure of the program file read from the. Further, it is connected to a personal computer 5 to be operated via an interface 61 forming a part of the communication means 6, and transmits and receives various signals and data to and from the personal computer 5.

The image pickup section 20 picks up an image of an operator's face as an object to be determined.
It includes a camera control unit 23, an A / D converter 24, and an interface 25. The light receiving element 22 is, for example, a CCD, and converts a light image incident through the lens 21 into an electric signal. The camera control unit 23 adjusts the imaging condition by adjusting the gain of the gain control for changing the amount of received light, adjusting the shutter speed, and the like, in response to the instruction from the signal control unit 10, and adjusts the light receiving element 22. Converts the converted electric signal into an A / D converter 24
And output via the interface 25.

The image calculation section 30 functions as an image processing section for performing a predetermined image processing, which will be described later, on the subject image picked up by the image pickup section 20, and the current image picked up by the image pickup section 20 A collation processing unit that performs a collation process between the image data of the target and the image data recorded in a registration dictionary 421 described later, and a determination unit that performs a permission / prohibition determination process of the operation target (the personal computer 5) based on the collation result. Function as The image processing, collation processing, and permission / refusal determination processing performed in the image calculation unit 30 will be described later in detail together with the description of each processing flow.

The work memory section 41 is a work memory section used for each processing performed by the image calculation section 30, and is constituted by a RAM capable of coping with frequent rewriting of data contents. In the continuous collation processing described later, an area for the continuous collation work memory 411 is secured in the work memory unit 41. The memory for continuous collation work 411 will be described later in detail with reference to FIG.

The data holding memory section 42 is a memory section for storing information required to hold the stored contents for a predetermined period, and is capable of rewriting the data contents and storing the data contents without using a backup power supply or the like. It is composed of a flash ROM that can hold the contents. In the data holding memory unit 42, a registration dictionary 421 for recording image data and the like of an authorized operator is formed, and the camera device (determination device main body) 1
An ID data memory 422 used to determine whether or not is a legitimate camera device 1 having the authority to determine whether or not to operate the personal computer (operation target) is formed. The registration dictionary 421 will be described later in detail with reference to FIG. Also, I
The D data memory will be described later in detail together with the description of the basic flow and the like.

As described above, the program memory section 11 is a memory section for recording the procedures of various processes in the camera device 1, and is an R section whose data contents are not erased.
OM.

On the other hand, the personal computer (operation target) 5 includes a signal control unit 50, a program memory unit 51, a data holding memory unit 52, an output processing unit 53, a flow program creation processing unit 54, a connected device drive processing unit 55, and an external device. Interface 62 to 59 and the above-described interface 62 connected to the camera device 1.

The signal control unit 50 serves as a central control unit for controlling a series of operations in the personal computer 5 to be operated. For this reason, the signal control unit 50 is connected to each unit of the personal computer 5, and the program memory unit 5
These operations are controlled in accordance with the procedure of the program file and the like read from Step 1. Further, it is connected to the camera device 1 which is the main body of the determination device via an interface 62 forming a part of the communication means 6, and sends and receives various signals and data to and from the camera device 1. Also, a power switch 70, a keyboard 71, a mouse 7 via the interfaces 56 to 59.
2. Speaker 73, monitor 74, printer 75, Internet 76, file recording device 77, and CD-R
It is connected to external devices such as the OM reader 78 and transmits and receives various signals and data to and from these external devices.

As described above, the program memory section 51 is a memory section for recording various processing procedures in the personal computer 5, and is constituted by a hard disk device. It should be noted that the program memory unit 51 is
-It may be configured by reading a program file or the like from an external device such as the ROM reader 78 into a RAM or the like as necessary.

The data holding memory section 52 is a memory section capable of holding recorded contents for a predetermined period, and is configured as a storage area secured in the same hard disk device as the program memory section 51. In the data holding memory unit 52, a file name or an external device name to which access is permitted according to an operator who has logged in to the personal computer 5 is recorded. The ID data used to determine whether or not the camera device 1 is authorized to perform the operation permission / non-permission determination is recorded. The data holding memory unit 52 may be constituted by a flash ROM or the like.

The output processing unit 53 and the flow program creation processing unit 54 create signals and data to be transmitted to the camera device 1 based on the instruction of the signal control unit 50.
Process.

The connected device drive processing unit 55, based on the instruction of the signal control unit 50, reads the CD-ROM
The control information for driving an external device such as 8 is created.

Next, the processing performed in the operation permission / refusal determination device having the above-described configuration will be described with reference to the flowcharts of FIGS.

FIGS. 3 and 4 show the entire flow of the processing performed in the present apparatus as a basic flow, which is divided into two figures for convenience of drawing.

Note that this basic flow is controlled by the signal control unit 50 of the personal computer 5 with the initiative in principle, and each flow that shifts from this basic flow is basically processed by the signal control unit 10 of the camera device 1 Take control with initiative.

In this basic flow, first, the camera device 1 and the personal computer 5 determine whether the ID has data (identification data) (step S10). Specifically, the signal control unit 10 of the camera device 1 checks whether ID data is recorded in the ID data memory 521 of the data holding memory unit 52, and the signal control unit 10 of the camera device 1 It is checked whether or not ID data is recorded in the ID data memory 422 of the unit 42. The ID data on the camera device 1 side is transmitted to the communication unit 6.
Is transmitted to the personal computer 5 side, and the signal control unit 50 of the personal computer 5 determines whether the ID data is recorded in the camera device 1 and the personal computer 5.

If no ID data is recorded in the camera device 1 and the personal computer 5 (NO in step S10),
It is determined that the camera device 1 has been connected to the personal computer 2 for the first time, and the process proceeds to an initial setting flow (step S12) including a process of giving ID data to the camera device 1 and the personal computer 2 for recording. This initial setting flow will be described later.

If the ID data is recorded in the camera device 1 and the personal computer 2 (YES in step S10), it is determined whether the ID data of the camera device 1 and the personal computer 5 match (step S14). That is, whether or not the camera device 1 is the genuine determination device main body having the authority to perform the operation permission / non-permission determination for the personal computer 5 to be operated is determined whether or not the camera device 1 has the same ID data as the personal computer 5. Is determined by

If the ID data of the camera device 1 and the ID data of the personal computer 5 do not match (NO in step S14), it is determined that the camera device 1 without authority to perform the operation permission determination of the personal computer 5 to be operated is connected. After that, the operator is informed that the ID data of the camera device 1 does not match the ID data of the personal computer 5 by displaying it on the monitor 74 or by sound from the speaker 73 (step S1).
6), end this basic flow. That is, login to the personal computer 5 by the camera device 1 is prohibited.

If the ID data of the camera device 1 matches the ID data of the personal computer 5 (YES in step S14), it is determined that the camera device 1 having the authority to determine whether or not to operate the personal computer 5 to be operated is connected. Is done. Then, the operator is asked whether or not to delete the ID data of the camera device 1 and the personal computer 5 (step S1).
8) If the operator instructs to perform the deletion process (YES in step S18), the process proceeds to an initialization flow (step S20) for deleting the ID data of the camera device 1 and the personal computer 5. This initialization flow is
An object of the present invention is to delete the ID data of the camera device 1 and the personal computer 5 so that the association between the two can be eliminated, and thereby the personal computer 5 to be operated can be connected to the camera device 1 as a new determination device main body. The specific procedure will be described later.

When ID data of camera device 1 and personal computer 5 is not deleted (NO in step S18)
Asks the operator whether to newly create or delete the registered dictionary 421 (step S22).
21 to be newly created or deleted (step S
22; YES), registration dictionary creation flow (step S2
Go to 4). This registration dictionary creation flow is a process performed by the signal processing unit 10 of the camera device 1 with the initiative.
The detailed procedure will be described later.

If the registration dictionary 421 is not newly created or deleted (NO in step S22), the flow shifts to the initial collation processing flow (step S26), in which the operator is an authorized operator registered in the registration dictionary 421. It is determined whether there is. This initial collation process flow is a process performed by the signal processing unit 10 of the camera device 1 under the initiative, and when a series of processes is completed, a determination result as to whether or not to permit login from the camera device 1 to the personal computer 5 is made. Sent. The detailed procedure will be described later.

If the login is not permitted through the initial collation processing flow (NO in step S28), the login to the personal computer 5, that is, the operation of the operation target is rejected. Specifically, the operation of the personal computer 5 is locked in order to refuse the operation of the personal computer 5 by the operator (step S
30), returning to the initial collation processing flow (step S26) again, and waiting for the login of the regular operator.

If login is permitted (step S28)
YES), indicating that the operator attempting to log in is determined to be a legitimate operator, and logs in to the personal computer 5 (step S32). If the operation of the personal computer 5 is already locked, this lock is released.

When the operator logs in to the personal computer 5 in this manner, the flow shifts to a continuous collation processing flow for determining whether or not the logged-in operator remains in front of the personal computer 5 (step S34). This continuous collation processing flow is a process performed by the camera device 1 with the initiative, and the result of the continuous collation is appropriate (the operator is in front of the personal computer).
As long as it is determined, the repetition is continuously performed, and when the continuous collation result is determined to be inconsistent (mismatch), the collation result is transmitted from the camera device 1 to the personal computer 5 for the first time (step S36). Therefore, the operator who has once logged in to the personal computer 5 can continuously operate and operate the personal computer 5 as long as the operator remains in front of the personal computer 5 because the continuous collation result is determined to be appropriate. Further, since this continuous collation process flow is a process performed by the camera apparatus 1 under the initiative, the personal computer 5 does not have the burden of the continuous collation process, and the operator continuously repeats the continuous collation process flow. During the operation, the operation of the personal computer 5 can be performed comfortably. The specific procedure of the continuous matching processing flow will be described later in detail.

If the result of the continuous collation is determined to be inconsistent (step S36), it indicates that the operator has left the personal computer 5, and the operation of the personal computer 5 is locked (step S38). Then, the process proceeds to a return waiting process flow (step S40) for determining whether or not the user has returned to the personal computer 5. This return wait processing flow is also performed by the camera apparatus 1 with the initiative, and the collation result is transmitted from the camera apparatus 1 to the personal computer 5 when a series of processing is completed. The detailed procedure will be described later.

In this return waiting processing flow, if it is determined that the operator has returned to the personal computer 5 (YES in step S42), the lock of the personal computer 5 is released (step S44), and the continuous collation processing flow is performed again. Returning to (Step S34), the returned operator checks the personal computer 5
Can be operated and operated.

On the other hand, if it is determined in this return waiting flow that the operator has not returned to the personal computer 5 (NO in step S42), the power switch 70 of the personal computer 5 or the personal computer 5 and the camera device 1 It is checked whether or not the communication cable 60 between them has been interrupted. If the communication cable 60 has not been interrupted (NO in step S46), the process returns to the return waiting process flow (step S40) and repeats the return wait process flow. On the other hand, if the power supply or communication has been interrupted (YES in step S46), the basic flow ends.

Hereinafter, each flow to be shifted from the basic flow will be sequentially described with reference to the drawings.

FIG. 5 shows an initial setting flow.

As described above, this initial setting flow is as follows.
This is executed when the ID data is not recorded in the camera device 1 and the personal computer 5 (NO in step S10 of the basic flow). The communication conditions between the PC and the personal computer 5 are set.

That is, in this initial setting flow, first, a predetermined CD is stored in the CD-ROM reader 78 of the personal computer 5.
It is determined whether a ROM is mounted (step S5)
0), when not mounted (NO in step S50)
In step S52, a notification to alert the operator to mount the CD-ROM is performed. Such notification can be made by the display of the monitor 74 or the sound of the speaker 73.

When the CD-ROM is loaded, communication conditions between the camera device 1 and the personal computer 5 are set (step S54). This communication condition setting is for setting a communication speed, a communication protocol, and the like, and may be performed automatically between the camera device 1 and the personal computer 5 or may be performed by an operator.

When the communication conditions are set in this way, the CD-
The ID data recorded in the ROM is read into the personal computer 5 (step S56), and the read ID data is written into the ID data memory 521 in the data holding memory unit 52 of the personal computer 5 (step S5).
8), transmitted to the camera device 1 via the communication means 6 and written into the ID data memory 422 in the data holding memory unit 42 of the camera device 1 (step S60), and the initial setting flow ends.

FIG. 6 shows an initialization flow.

As described above, this initialization flow is performed when it is confirmed that the same ID data is recorded in the camera device 1 and the personal computer 5. When the user instructs to delete the ID data recorded in (1) and cancel the association between them (YE in step S18 of the basic flow)
This is executed in S).

In this initialization flow, a predetermined CD-RO is first stored in the CD-ROM reader 78 of the personal computer 5 again.
It is determined whether M is mounted (step S70),
If the CD-ROM has not been mounted (NO in step S70), a notification to alert the operator to mount the CD-ROM is performed (step S72).

The reason why the CD-ROM is required to be mounted in this initialization flow is to use the CD-ROM as a key for the initialization process. Even if the CD-ROM is mounted, the IDs of the personal computer and the CD-ROM can be obtained. If the data does not match (NO in step S74), a notification to call for the installation of a regular CD-ROM is issued (step S72).

If a CD-ROM having ID data matching the ID data recorded in the personal computer 5 is loaded (YES in step S74), the data stored in the data holding memory sections 42 and 52 of the personal computer 1 and the camera device 5 are stored. The ID data recorded in the ID data memories 422 and 521 are erased (steps S76 and S78), and the initialization flow ends.

FIGS. 7 and 8 show a registration dictionary creation flow.

As described above, this registered dictionary creation flow is executed when a new creation or deletion of the registered dictionary 421 is instructed by the operator.

The registration dictionary 421 is for registering information of an operator to be permitted to operate and operate the personal computer 5 to be operated. Therefore, first, items to be registered in the registered dictionary 421 and the data configuration of the registered dictionary 421 will be described with reference to FIG.

The registration dictionary 421 in this embodiment is
Up to eight operators can be registered, and three variations of image data can be registered for each registered operator. Each registrant is represented by a registrant number n of 1 to 8 for convenience, and each variation is represented by a variation number m of 1 to 3 for convenience. Note that the variation number m
The image data of = 1 is used in the initial collation processing (first collation processing for determining whether to permit the operator to log in) described later, and the variation number m =
The image data of 2, 3 is the variation number m = 1.
Along with the image data, a continuous collation process described later (a collation process that is performed continuously to determine whether or not the operator remains in front of the personal computer after login is once permitted) and a return waiting process (from the front of the personal computer) This is used in a collation process for determining whether or not a distant operator has returned.

The average illuminance value is recorded in each image data (n, m). The illuminance average value is obtained by averaging the illuminance of the entire captured image before the image processing flow described later is performed and converted into image data of a predetermined format. Used for processing.

Further, for each registrant n = 1 to 8, the name, the initial verification collation rate, the continuous collation collation rate, and the rank are recorded. The rank of the registrant is obtained by classifying the access authority to the confidential data and the like in the personal computer 5, and is expressed as, for example, A rank, B rank, and C rank as shown in FIG. Such a ranking is merely an example, and if it is an in-house network system or the like, a department or the like may be used as the rank. The collation rate at the time of initial collation is a collation rate serving as a criterion for determining the suitability of the collation in the initial collation processing described later. Is the matching rate. For example, in the initial collation processing, if the collation rate between the currently captured image data and the registered image data exceeds the collation rate at the time of the initial collation, the initial collation result is determined to be appropriate. In the present embodiment, the initial verification collation rate is also used in a return waiting process to be described later.

Returning to FIGS. 7 and 8, each procedure of a registration dictionary creation flow for registering a new regular operator in such a registered dictionary 421 or deleting a registered regular operator will be described.

In this registration dictionary creation flow, first, a predetermined CD-ROM is stored in the CD-ROM reader 78 of the personal computer 5.
It is determined whether the ROM is mounted (step S8).
0) If not mounted, CD
-Notification to evoke ROM installation is performed (step S8)
2).

In the registration dictionary creation flow, the CD-R
The request for mounting the OM is to use the CD-ROM as a key for the registration dictionary creation processing and the deletion processing. For this reason, even if the CD-ROM is mounted, it must match the ID data of the personal computer 5. (Step S84
(NO at step S82), a notification for invoking the installation of a regular CD-ROM is performed (step S82).

When a CD-ROM having ID data matching the ID data of the personal computer 5 is loaded (YES in step S84), the registration dictionary 421 is sent to the operator.
(Step S86). The user is asked whether to create a new dictionary or to delete the registered dictionary 421.

If the registration dictionary 421 is to be deleted (deletion in step S86), the operator is prompted to input a registrant number n or a registrant name to be deleted (step S86).
88), the input registrant number or registrant registration data is deleted from the registration dictionary 421 (step S9).
0). After performing a predetermined post-deletion process such as filling in data of a registrant of a registrant number that is lower than the registrant that has just been deleted (step S92), the personal computer 5 informs the user that the registration dictionary creation flow is to be terminated. (Step S12)
0), the registration dictionary creation flow ends.

On the other hand, if the registration dictionary 421 is to be newly created (newly created in step S86), the registration dictionary 42
The registration dictionary 42 depends on whether or not the maximum number of registered registrant numbers is smaller than eight, which is the maximum number of registered persons.
It is determined whether or not there is a free space in Step 1 (Step S9).
4).

If there is no free space in the registered dictionary 421 (NO in step S94), the operator is notified that the registered dictionary 421 is full (step S96), and the completion of the registered dictionary creation flow is transmitted to the personal computer 5. From (Step S12
0), the registration dictionary creation flow ends.

If there is a vacancy in the registration dictionary 421 (YES in step S94), the newly created registrant number n is set as a number obtained by adding 1 to the already registered largest registrant number (step S98).

Subsequently, the operator is prompted to input a name, a log-in collation rate, a continuous collation collation rate, a rank, and the like as registrant information on a registrant to be newly registered, and register the entered registrant information. The information is recorded in the dictionary 421 (step S100).

When the registrant information is input, the process proceeds to a process for creating registrant image data. In this embodiment, as described above, in order to register image data of three variations having different postures for each registrant, the variation number m is first initialized to 1 (step S10).
2).

Subsequently, an image of the operator is actually taken (step S104). The shooting of the registered image will be described with reference to FIG. 9 as an imaging processing flow.

In this imaging processing flow, first, the operator who is a new registrant is notified of the photographing and the posture to be taken by the operator by a screen display on the monitor 74 or by a sound from the speaker 73 ( Step S13
0), an image is taken by the imaging unit 20, and an image of the operator who is a new registrant is taken into the working memory unit 41 via the image calculation unit 30 (step S132). And
The image captured by the image calculation unit 30 is reduced to a predetermined size (step S134), and the reduced image is subjected to optimal density conversion to obtain a captured image (step S1).
36). Note that the optimal density conversion is image processing for improving the contrast of an image by redeploying the density range (effective gradation) of the captured image to a maximum gradation width of 0 to 127, for example.

Returning again to the registration dictionary creation flow of FIG.
The image captured by the imaging processing flow is displayed on the monitor 74 (step S106), and the operator who is a new registrant is asked whether or not this image is to be a registered image (step S108). On the monitor 74, markers 81 are displayed in an effective image area 80 of the captured image as shown in FIG. The markers 81 indicate the four corners of an important area 82 that is particularly important in the effective image area in the image processing described later. For this reason, it is desirable that the operator's face area, which is a characteristic part of the object to be determined, be imaged so as to be exactly within the emphasis area 82 indicated by the markers 81. Therefore, the operator not only confirms the facial expression and the like that has been photographed, but also confirms whether or not the face portion is within the important area 82, and instructs re-photographing if there is any inconvenience (step S).
108 to re-shooting). If re-shooting is instructed, shooting of the registered image is performed again by the shooting processing flow.

On the other hand, if the operator accepts that this image is a registered image (OK in step S108), the illuminance average value of this image is obtained, and the illuminance average value thus obtained is stored in the registration dictionary 421. The illuminant is recorded as the average illuminance value of the image data of the registrant n and the variation number m (step S110). That is, the average illuminance is obtained by simply averaging the illuminance of the entire region of the image before converting the image to be registered into image data of a predetermined format, without biasing a part of the entire region of the image. Value.

Subsequently, the photographed image is subjected to image processing to create image data of a predetermined format (step S1).
12). This image processing is performed as an image processing flow shown in FIG.
20 will be described with reference to FIG.

This image processing flow is performed when the image calculation section 30 functions as an image processing section.

In this image processing flow, the emphasis area 82 is set so as to include the face from the effective image area 80 of the image including the face as shown in FIG.
Is determined (step S140). This emphasis area 82
As described as image processing in another embodiment described later, the position is moved in the effective image area 80, the size is enlarged / reduced, or the shape is deformed, The face can be made to fall exactly in the important area 82.

When the emphasis area 82 is determined, FIG.
The effective image area 80 is divided into a plurality of blocks 84 as shown in FIG.
, 85 ... (step S142). At this time, the blocks 84 in the important area 82 are divided so as to have a smaller area than the blocks 85 in the other areas. Here, the entire effective image area is divided into 9 × 7 areas, and the important area 82 arranged at the substantially central portion is divided into 6 × 6 blocks 85. Each of the blocks 84 has an area 1/4 that of the blocks 85 in the other area. The size (area) and the number of each block are determined by the image processing capability of the image processing unit and the collation determination accuracy described later. Can be arbitrarily set in accordance with a request or the like, and it is sufficient that the image is divided so as to capture the features of the image.

When the effective image area 80 of the image is divided into a plurality of blocks 84, 85,..., The density (brightness) representative of each of the divided blocks 84, 85,. The density value is determined as a block characteristic value (step S14).
4). As the block characteristic value, various values can be adopted as long as the value can be uniquely determined as a value representative of each block. For example, the density (brightness) of all pixels included in each block is used.
An average value, a frequently occurring density value or color data among the densities of all pixels can be used.

When the block characteristic values of the blocks 84... 85 are determined, image data of the image is created as a set of the block characteristic values (step S).
146). In other words, this image data gives a relatively large amount of information to the characteristic portion of the object to be determined identified as the important region 82 in the image captured by the imaging unit 20 to reduce the weight occupying the image data. It has been created by raising it.

Returning again to the registration dictionary creation flow of FIG.
The image data created by the image processing flow is recorded as registered image data of the image variation number (m) of the registrant (n) of the registered dictionary (step S114).

When the registration of one image variation is completed, whether or not registration of all (three) image variations is completed is determined by whether or not m used as a variation number counter is 3. (Step S116), if only the first or second image variation has been registered (NO in step S116), the variation number m is incremented (m = m + 1; step S11).
8), and return to the imaging and registration processing for the next image variation (steps S104 to S116).

If the registration of all the image variations has been completed (YES in step S116), the fact that registration dictionary creation has been completed is transmitted to the personal computer (step S120).
The registration dictionary creation flow ends.

FIGS. 11 and 12 show the initial collation processing flow.

As described above, this initial collation processing flow is to collate whether or not the operator is a legitimate operator registered in the registration dictionary 421 and decide whether or not to permit login. is there. In this initial collation processing flow, the current image data obtained by photographing the operator who is going to log in is one of three registered image data recorded for each registrant registered in the registration dictionary 421. Whether or not to permit the operator to log in is determined based on whether or not the image data matches the registered image data of variation number m = 1 photographed face-on.

In the initial collation processing flow, first, an image of the operator in front of the personal computer 5 is photographed as a current image (step S160). The shooting processing flow shown in FIG. 9 described above is applied to the shooting of the current image.

The current image thus captured is displayed on the monitor 74.
Is displayed (step S162), and the operator who logs in is asked whether or not to perform the initial collation processing with this image (step S164). At this time, the marker 81 indicating the important area 82 is superimposed on the image in the monitor 74.
.. Are displayed (see FIG. 19), and the operator
Check if the face is within 2. If re-shooting is instructed by the operator (re-shooting in step S164), the process returns to shooting of the current image (step S160), and shooting of the current image is performed again.

If the photographed current image is accepted by the operator (OK in step S164), the average illuminance of the photographed current image is obtained (step S166). The illuminance average value of the current image is used when comparing the current image with registered image data recorded in the registered dictionary 421.
In order to improve the accuracy of this collation, it is used in a brightness correction process for unifying the brightness conditions at the time of photographing both images and the density conditions after optimal density conversion.

In the initial collation processing, the current image thus photographed is collated with the image data registered in the registration dictionary 421. Since the image data to be collated with the current image is sequentially called from the registration dictionary 421, the registration is not performed here. The user number n and the variation number m are used as counters. Therefore, first, both the counters n and m are initialized to 1 (step S168). In the following, the images of the registrant number n and the variation number m in the registration dictionary 421 are abbreviated as registered images (n, m), and the registrant information of the registrant number n is compared, for example, at the time of initial collation. Abbreviated as rate (n).

Then, the illuminance average value (n, m) of the registered image data (n, m) is called from the registration dictionary 421 (step S170), and the illuminance average value (n, m) and the current image that has just been shot are displayed. Is obtained (step S172), and the brightness correction processing of the current image taken is performed according to the difference between the obtained illuminance average values (step S174). This luminance correction processing is to remove a difference in brightness condition at the time of shooting between the current image and the registered image (n, m), a difference in density condition after the optimal density conversion, and the like. The brightness is uniformly increased or decreased to realize accurate collation of both images.

The current image whose luminance has been corrected in this way is subjected to image processing, and current image data of a predetermined format is created (step S176). In this image processing, the image processing flow of FIG. 10 described above is applied, and image data in which the predetermined emphasis area 82 is emphasized is created.

Next, the registered image data (n, m) for collation with the current image data is called from the registration dictionary 421 and temporarily stored in the working memory unit 41 (step S178). Preliminary collation between the data and the registered image data (n, m) is performed (step S17).
9). The preliminary collation processing is performed by the image calculation unit 30 functioning as a collation processing unit. The preliminary collation processing is performed for each of the blocks constituting the current image data and the registered image data (n, m). This is done by comparing the block characteristic values of the blocks. Any number of blocks may be used for the preliminary collation processing. For example, blocks 851..., 852... Around the effective image area shown in FIG. Block 84 in the important area 82 shown in c)
1 ... can be used.

The proportion of blocks that have been matched by the preliminary collation is determined as the collation rate. This collation rate is compared with a predetermined preliminary collation rate (step S180), and if it is lower than the preliminary collation rate (NO in step S180). ), It is determined that the current image data and the registered image data (n, m) are not suitable without performing comparison and collation of all blocks.

On the other hand, if the ratio exceeds the preliminary collation rate (YES in step S180), it is determined that the collation between the current image data and the registered image data (n, m) is possible, and all blocks are compared. Collated (step S
181). This comparison and matching process is also a process performed by the image calculation unit 30 functioning as a matching processing unit.
More specifically, this is performed by comparing the block characteristic values of the corresponding blocks of the mosaiced current image data and registered image data. That is, the blocks 84 in the important area 82 of the current image data ...
Is the corresponding block 84 of the registered image data (n, m).
... and the blocks 85 in other areas of the current image data
.. Correspond to the corresponding blocks 8 of the registered image data (n, m).
5 are compared with each other. Then, the ratio of the number of matching blocks is obtained as the matching rate. Therefore,
The emphasis area 82 has a larger number of blocks per predetermined area than the other areas, so that the emphasis area 8 occupies the collation rate of the entire image.
The weight of No. 2 is larger than the other areas, and as a result, the weight is more important than the other areas.

If the collation rate between the current image and the registered image (n, m) is obtained in this way, the collation rate (n) at the time of initial collation of the registrant of the registrant number n who has collated is now registered in the registration dictionary 42.
1 (step S182), and it is determined whether the collation rate between the current image data and the registered image data (n, m) exceeds the initial collation collation rate (n) (step S184). This determination process is also performed when the image calculation unit 30 functions as a determination unit.

The result of the preliminary collation is inconsistent (NO in step S180) or the current image and the registered image (n,
If the collation rate of m) does not exceed the collation rate (n) at the time of the initial collation, that is, if the collation result is not suitable (step S1)
NO at 84), all registrants 8 registered in the registration dictionary 421
It is determined whether or not the verification of the person has been completed based on whether or not the registrant number n used as the counter is 8 (step S186). If the verification of all the registrants has not been completed yet (step S186) (NO at S186), increments this registrant number n (step S188),
The process returns to a series of processes (steps S170 to S184) for collating the photographed current image data with the next registrant.

If the collation of all registrants has been completed (YES in step S 186), it indicates that the operator who is going to log in is not a legitimate operator registered in registration dictionary 421. The fact that there is a collation error and the collation rate between the current image data and each registrant n = 1 to 8 registered in the registration dictionary 421 are transmitted to the personal computer 5 as collation rate data for each registrant (step S19).
0), the initial collation processing flow ends.

On the other hand, if the collation rate between the current image data of the operator who is now logging in and the registered image data (n, m) is higher than the initial collation rate (n) (YES in step S184), The operator who is about to log in is determined to be a regular operator registered as the registrant number n in the registration dictionary 421.

Therefore, the current image data that has been photographed and whose collation result is determined to be appropriate is recorded in the registration dictionary 421 as the latest image data of the registrant registered with the registrant number n in the registration dictionary 421. Update the content. Specifically, the current image data is overwritten as the image data (n, m) of the registered dictionary 421, and the average illuminance value of the current image is averaged over the illuminance average of the registered image (n, m) of the registered dictionary 421. The value is overwritten as a value (step S192).

Then, along with the fact that the initial collation is determined to be conforming, the initial collation result is determined to be conforming, and the registrant information of the operator (regular operator with registrant number n) to be logged in is read from the registration dictionary 421. To the personal computer 5 (step S194), and the initial collation processing flow ends. At this time, the contents transmitted to the personal computer 5 include the name and rank of this operator (regular operator having the registrant number n).

The personal computer 5 to be actuated logs in the operator who is the authorized registrant according to the judgment result of the initial collation transmitted from the camera device 1 which is the main body of the judgment device, and logs in the transmitted data. The access right to the confidential file and the like is managed according to the rank of the operator.

FIG. 13 and FIG. 14 show the flow of the continuous collation processing.

As described above, in the continuous collation processing flow, once the operator has logged in to the personal computer 5, it is determined whether or not the operator remains in front of the personal computer 5. More specifically, a current image of an area in front of the personal computer 5 where the operator should be present is stored in the working memory unit 41 of the camera device 1 for the continuous collation work memory 4.
The image calculation unit 30 of the camera device 1 determines whether or not the image matches any one of the three variations of the immediately preceding image recorded in 11.

The memory for continuous collation work 411 has the structure shown in FIG.
As shown in FIG. 8, three types of image data of a worker with a registrant number n who is logged in are represented by a variation number m =
1 to 3 are recorded as the immediately preceding image data, and the collation rate with the current image data is recorded up to 10 times for each immediately preceding image data = 1 to 3.

As described above, the continuous collation work memory 4
The reason why the collation rate for 10 times is recorded in 11 is that, in this continuous collation processing, the current image taken is 3
Even if the image does not match any of the images immediately before the variation, the current image is repeatedly imaged 10 times, and is continued only when the matching result does not match any of the current images captured 10 times again. This is to record the results of the ten times of collation in order to determine that the collation result is not suitable.

The reason why the collation operation is repeated ten times in the continuous collation processing is that the operator logs in to the personal computer 5 at the time of the continuous collation and is working. May have a different posture from any of the three types of previous image data,
In such a case, it is inappropriate to immediately determine that the continuous collation result is inconsistent and prohibit the operation of the personal computer 5.

Returning to FIG. 13 and FIG. 14, in this continuous collation processing flow, first, three registered images of the variation number m = 1 to 3 of the normal operator of the registrant number n who has logged in through the initial collation processing flow. Data (n, 1),
(N, 2), (n, 3) are called, and these are stored as the immediately preceding image data (1), (2), and (3) of the variation numbers m = 1 to 3, respectively.
11 is stored. That is, the image data registered in the registration dictionary 421 is used as initial data of the immediately preceding image data used for the continuous matching process.

Subsequently, the collation rate (n) at the time of continuous collation of the registrant n who is currently logged in is called from the registration dictionary 421 (step S202), and thereafter, the process proceeds to the repeated collation operation.

In this continuous collation operation, first, a collation trial counter c for counting the number of collations repeated up to 10 times is initialized to 1 (step S204).

Then, the personal computer 5 where the operator should be present
Is taken as the current image (step S20).
6). The shooting processing flow shown in FIG. 9 described above is applied to the shooting of the current image. Note that, as described above, since the continuous collation processing is performed in the background without the intervention of the personal computer 5 which is the operation target, the confirmation display of the current image captured to the operator is not performed.

The photographed current image is subjected to image processing to create current image data in a predetermined format (step S20).
8). For this image processing, the image processing flow shown in FIG. 10 described above is applied.

The current image data thus created is first stored in the first memory 411 for continuous collation work.
Comparison and collation are performed with the image data (1) immediately before the variation, and the collation rate is recorded in the continuous collation work memory 411 as the collation rate data (c, 1) with the collation trial counter c = 1 with the variation number m = 1. (Step S21
0). Note that the comparison and comparison between the current image data and the immediately preceding image data are performed in the same manner as the above-described comparison processing between the current image data and the registered image data in the initial matching process. It is determined whether or not they match, and the number of matching blocks out of all blocks is used as the matching rate.

Then, it is determined whether the collation rate between the current image data and the immediately preceding image data (1) thus determined is greater than the collation rate during continuous collation (n) (step S2).
12) If it exceeds (YE in step S212)
S) It is determined that the regular operator with the registrant number n is present in front of the personal computer 5 in the posture of the immediately preceding image data (1), and the continuous collation result is appropriate.

Therefore, the current image data which has just been photographed and whose continuation collation result is determined to be suitable is stored in the continuation collation work memory 411 in the immediately preceding image data (1) of the first variation.
And updates the recorded content of the continuous collation work memory 411 (step S214). After updating the immediately preceding image data in this way, the collation trial counter c is initialized to 1, the process returns to capturing a new current image, and this continuous collation processing is repeated.

On the other hand, if the collation rate between the current image data and the immediately preceding image data (1) is lower than the collation rate during continuous collation (n) (NO in step S212), the continuous collation with the current image data is continued. The comparison with the immediately preceding image data (2) of the second variation recorded in the working memory 411 is performed. Then, the collation rate obtained by the comparison collation is recorded in the continuous collation work memory 411 as the collation rate data (c, 2) of the collation trial counter c of the variation number m = 2 (step S216).

Then, it is determined whether the collation rate between the current image data and the immediately preceding image data (2) is greater than the collation rate during continuous collation (n) (step S218), and if it is greater (step S218). YES), it is determined that the normal operator of the registrant number n is present in front of the personal computer in the posture of the immediately preceding image data (2), and the continuous collation result is appropriate.

Therefore, the current image data which has just been photographed and whose continuation collation result is determined to be appropriate is stored in the continuation collation work memory 411 in the immediately preceding image data (2) of the second variation.
To update the recorded content of the continuous collation work memory 411 (step S220). After updating the immediately preceding image data in this way, the collation trial counter c is initialized to 1, the process returns to capturing a new current image, and this continuous collation processing is repeated.

On the other hand, if the collation rate between the current image data and the immediately preceding image data (2) is lower than the collation rate during continuous collation (n) (NO in step S218), then the continuous collation with the current image data is performed. The comparison with the immediately preceding image data (3) of the variation number m = 3 in the working memory 411 is performed. Then, the collation rate obtained by the comparison collation is recorded in the continuous collation work memory 411 as the collation rate data (c, 3) of the collation trial counter c of the variation number m = 3 (step S222).

Then, it is determined whether or not the collation rate between the current image data and the immediately preceding image data (3) exceeds the collation rate during continuous collation (n) (step S224). YES), it is determined that the normal operator with the registrant number n is present in front of the personal computer in the posture of the immediately preceding image data (3), and the continuous collation result is appropriate.

Therefore, the current image data which has just been photographed and whose continuous collation result is determined to be suitable is stored in the continuous collation work memory 411 in the immediately preceding image data (3) of the third variation.
To update the recorded contents of the continuous collation work memory 411 (step S226). After updating the immediately preceding image data in this way, the collation trial counter c is initialized to 1, the process returns to capturing a new current image, and this continuous collation processing is repeated.

As a result of the comparison between the current image data and the three immediately preceding image data m = 1 to 3, if the current image data does not match any of the three immediately preceding image data (steps S212, S218, and S224). Both are N
O) It is determined whether or not the number of collation trials is the tenth (step S228). If the number of collation trials is not the tenth (that is, if it is the ninth or less), the collation trial counter c is incremented. (Step S23
0), the process returns to the shooting of the new current image, and the comparison and comparison between the newly shot current image and the immediately preceding images of the three variations are repeated.

On the other hand, if the number of collation trials is the tenth (YES in step S 228), it is determined that the continuous collation is finally inappropriate, and the immediately preceding images of the three variations recorded in the continuous collation work memory 411. An average value of a total of 30 collation rates obtained by performing 10 times for each data is obtained (step S232), and the fact that there is a continuous collation error and the average value of the 30 collation rates are transmitted to the personal computer 5 ( Step S234), the continuation collation processing flow ends.

In this manner, the continuous collation result is determined to be inconsistent (error), and the personal computer 5 which is the operation target having received the determination result is determined to have left the operator in front of the personal computer 5. 5 is locked, and the transmitted average value of the collation rate is displayed on the monitor 74. The average value of the collation rate is displayed so that the operator can check the collation rate.

FIG. 15 and FIG. 16 show the flow of the processing for waiting for return.

[0138] As described above, the return wait processing flow is such that when it is determined in the above continuous collation processing that the operator who has logged in to the personal computer 5 has left from the personal computer 5, the operator is in front of the personal computer 5. It is to repeatedly check whether or not it has returned.

In this return wait processing flow, the personal computer 5
Is compared with the image data of all variations m = 1 to 3 of all registrants n = 1 to 8 registered in the registration dictionary 421 and registered in the registration dictionary 421. It is determined whether any authorized operator has returned to the personal computer 5. In this way, all the registrants n = 1 to 8 registered in the registration dictionary 421 are to be compared because the registered regular operator has already logged in the personal computer 5 once. This is so that anyone can restart the operation of the personal computer 5 relatively easily. Also, variations m = 1 to 3 of all images of each registrant n = 1 to 8
Is set as the target of the collation, so that the operator who has returned to the personal computer 5 can restart the operation of the personal computer 5 relatively easily, compared with the initial collation of whether to permit the login. It is.

In the return waiting processing flow, first, the image before the personal computer 5 is photographed as the current image (step S240). The shooting processing flow shown in FIG. 9 described above is applied to the shooting of the current image. Then, the captured current image is displayed on the monitor 74 (step S24).
2). However, in this return wait processing flow, the operator is not confirmed whether to perform re-imaging. If the operator has not returned to the personal computer 5, it is meaningless, and even if the operator who has returned to the personal computer 5 is dissatisfied with performing the return waiting process on the current image, the above-described operation is performed. This is because the return waiting process is repeatedly performed, so that it is sufficient to wait for the next current image to be captured.

Subsequently, the illuminance average value of the photographed current image is obtained (step S244). The illuminance average value of the current image is used to improve the accuracy of the matching when the current image is compared with the registered image in the registered dictionary 421.
It is used for a brightness correction process for unifying brightness conditions at the time of photographing, density conditions after optimal density conversion, and the like.

In the return waiting process, the current image thus photographed is compared with the registered image data recorded in the registration dictionary 421. Since the registered image data to be compared with the current image is sequentially called from the registration dictionary 421, Then, the registrant number n and the variation number m are used as counters. Therefore, both the counters n and m are initialized to 1 (step S246).

Then, the illuminance average value (n, m) of the registered image data (n, m) is called from the registration dictionary 421 (step S248), and the illuminance average value (n, m) and the current image that has just been shot are displayed. Is obtained (step S250), and the luminance of the current image is corrected according to the difference between the obtained illuminance average values (step S25).
2).

The current image thus corrected for brightness is subjected to image processing to create current image data of a predetermined format (step S254). For this image processing, the image processing flow shown in FIG. 10 described above is applied.

Next, the registered image data (n, m) for collation with the current image data is called from the registered dictionary 421 and temporarily stored in the working memory unit 41 (step S256). Similar to the initial collation processing, preliminary collation between the current image data and the registered image data (n, m) is performed (step S257). If the collation rate obtained as a result of the preliminary collation is lower than the predetermined preliminary collation rate (NO in step S258), the current image data and the registered image data (n, m) are determined to be incompatible.

On the other hand, if the collation rate obtained as a result of the preliminary collation is higher than the preliminary collation rate (YE in step S258).
S), all blocks of the current image data and the registered image data (n, m) are compared and collated, and a collation rate is obtained (step S259).

If the collation rate between the current image data and the registered image data (n, m) is obtained in this way, the collation rate (n) at the time of initial collation of the registrant of the registrant number n who has collated the data is registered dictionary. 421 (step S260),
It is determined whether the collation rate between the current image data and the registered image data (n, m) exceeds the collation rate (n) at the time of the initial collation (step S262).

The current image data and the registered image data (n,
If the collation rate of m) exceeds the collation rate (n) at the time of the initial collation, the operator who has just returned to the personal computer 5 is a regular operator registered in the registration dictionary 421 with the registrant number n. Is determined (YES in step S262).

Therefore, the current image data that has just been photographed and whose collation result is determined to be compatible is recorded as the latest image of the authorized operator registered with the registrant number n in the registered dictionary 421, and the recorded contents of the registered dictionary 421 To update. More specifically, the current image data is overwritten as the registered image data of the registered image (n, m) of the registered dictionary 421, and the average illuminance value of the current image is similarly updated to the registered image (n, m) of the registered dictionary 421. Overwrite as the illuminance average value (step S
264).

Then, the registrant information of the authorized operator having the registrant number n returned to the personal computer 5 is read out from the registered dictionary 421 and transmitted to the personal computer 5 (step S).
266), the return waiting processing flow ends.

On the other hand, if the collation rate between the current image data and the registered image data (n, m) does not exceed the initial collation collation rate (n), that is, if the collation result is inappropriate (NO in step S262), the current It is determined whether or not the comparison of all the variations m = 1 to 3 has been completed for the registered registrant n (step S268). If the comparison of all the variations m = 1 to 3 has not been completed yet, (NO in step S268), increments the variation number m which is a counter (step S268).
270), and returns to a series of processes (steps S250 to S262) for collating the captured current image with the next registered image.

If the collation of all image variations m = 1 to 3 for the registrant of the registrant number n has been completed (YES in step S268), the collation of all registrants n = 1 to 8 has been completed. Is determined (step S272), and if the collation of all registrants n = 1 to 8 has not been completed yet (N in step S272)
O), the registrant number n which is a counter is incremented and the variation number m of the image is initialized to 1 (step S274), and a series of processes for collating the taken current image with the next registered image (steps S250 to S26)
Return to 2).

If the collation of the variations m = 1 to 3 of all the images of all registrants n = 1 to 8 has been completed (YES in step S272), the personal computer 5
It is determined that the regular operator registered in the registration dictionary 421 has not returned before the registration dictionary 421, and that there is a collation error, and for each registrant (n = 1 to 8), the variation of each image (m = 1 to The collation rate for each 3) is transmitted to the personal computer 5 (step S276), and the return wait flow ends.

According to such an operation permission / prohibition judging device, a characteristic portion of the object to be determined, such as the face of the operator, is specified as the important region 82 in the image picked up by the image pickup section 20. Is divided into smaller blocks of other areas, so that image data in which the image information in the important area 82 is more important than the image information in the other areas is created.
For this reason, the difference between the current image data and the registered image data in the important region 82 can be detected sensitively, and more accurate and accurate collation can be realized. In addition, other regions mainly including a background portion and the like can be realized. Since a slight difference between the two can be tolerated and a change in the imaging environment can be dealt with, erroneous determination can be prevented.

In the initial collation process and the return waiting process, before performing the comparison collation process for all the blocks constituting the image data, only some blocks of the current image data and the registered image data are compared. Preliminary collation to be collated is performed, and only when this preliminary collation result satisfies a predetermined preliminary collation determination standard, the comparison collation processing of all blocks is performed. For registered image data that is determined to be clearly incompatible with the current image data only, it is not necessary to perform comparison and collation of all blocks, so registered image data that should be collated with current image data can be narrowed down in advance. As a result, the time required for the collation determination processing can be reduced as a whole.

The important area 8 together with the captured image is also displayed.
2 is displayed on the monitor 74, the operator checks the position of the important area 82 in the captured image by looking at the monitor 74, and checks the current image so that the operator's face is included in the important area 82. Can be instructed again, and this ensures that the face, which is the characteristic part of the object to be determined, is placed in the focus area 82
And a more accurate and accurate collation can be realized.

Next, another embodiment will be described. In the following description, the same components as those in the above-described embodiment will not be described repeatedly, and only differences from the above-described embodiment will be described.

In this embodiment, in the image processing flow of the above embodiment (FIG. 10), the position of the important area 82 of the effective image area 80 of the captured image is moved, the size is enlarged or reduced, or The shape can be changed. Accordingly, the initial collation processing (FIG. 11, FIG.
In FIG. 12) and the like, there is a slight difference in the specific procedure of comparison and comparison between the current image data and the registered image data.

First, referring to FIG.
The movement of No. 2 will be described.

The movement of the emphasis area 82 is shown in FIG.
As shown in the figure, when the characteristic portion (the face portion of the operator) of the object to be determined is located at a biased position in the effective image region 80 of the captured image, the important region 82 is moved to this portion. The purpose is to include the characteristic portion of the object to be determined in the emphasis area 82.

The movement of the emphasis area 82 may be automatically performed by an operation permission / rejection determination device (specifically, the image processing unit 30 or the like) by various known optimization methods, or
The operator views the monitor 74 with the mouse or the like and focuses on the area 82.
May be performed, and may be performed based on this instruction. The same applies to enlargement / reduction and deformation of the important area 82 described later.

If the important area 82 is moved so as to include the characteristic portion of the object to be determined, the blocks 84 in the effective image area 82 are replaced with the blocks in other areas as shown in FIG. 85. The effective image area 80 is divided into a plurality of blocks 84, 85,. In a later process of collating the current image data with the registered image data, the corresponding blocks 84,.
85 in order to compare and match 85.
The number of divisions of 4... Is the same as the division processing (FIG. 20B) when the important area 82 is not moved (here, 6 × 6).
Number). Also, the division of the blocks 85... Of the other areas is desirably the same as that when the important area 82 is not moved (here, 9 × 7), except for the part hidden by the important area 82. .

When the effective image area 80 is divided into a plurality of blocks 84, 85,..., Block characteristic values are obtained for each of the blocks 84, 85,. Thus, image data as a set of the block characteristic values is obtained.

In this way, image data in which the emphasis area 82 has been moved is obtained.
(C)) shows that, as compared with the image data in which the emphasis area 82 has not moved (FIG. 20 (c)), information about some blocks 85... Missing (for example, coordinates (2, C) in the figure,
(2, D), etc., and the information of some blocks 85... Increases (for example, coordinates (6, C),
(6, D) etc.). Therefore, when such image data is compared and collated with other image data, the corresponding blocks 84... Are compared and collated with respect to the important area 82, but the corresponding blocks 84 are correlated with other areas. 8
5 are compared and compared, and the portions where the corresponding blocks 85 are not present are not compared and compared, and are not included in the calculation of the matching rate of the entire image.

By moving the focus area 82 in this way,
The position of the object to be determined or the characteristic portion of the object to be determined (here, the face portion of the operator) is not constant with respect to the imaging unit 20, and the effective image of the image of the object to be determined or the characteristic portion of the object to be determined is captured. Even if it is located at an uneven position in the area 80, it can be handled by image processing, so that particularly excellent operability can be obtained, and the emphasis is placed on the object to be determined or the characteristic portion of the object without re-photographing. Since it can be placed in the area 82, a series of operation times from imaging to collation determination can be reduced, and the object to be determined or the characteristic portion of the object to be determined can be reliably contained in the important area 82. Accurate collation determination can be performed.

Next, referring to FIG.
2 will be described.

The enlargement / reduction of the emphasis area 82 is shown in FIG.
As shown in (a), the effective image area 8 of the captured image
When the characteristic portion (the face portion of the operator) of the object to be determined is imaged in an excessively large or small value within 0, the emphasis area 82 is set according to the size of this characteristic part in the effective image area 80. It is intended to enlarge or reduce the size so that the characteristic portion of the object to be determined is just included in the important area 82.

When the important area 82 is enlarged or reduced, as shown in FIGS. 22 (b) and (c), the effective image area 80 is moved in the same manner as the movement of the important area 82. Are divided into a plurality of blocks 84... 85, and the block characteristic values are obtained to create image data.

By enlarging / reducing the important area 82 in this way, the distance of the object to be determined or the characteristic portion of the object to be determined to the image pickup unit 20 when the image is captured is not always constant, and the object to be determined or the object to be determined Even if the characteristic part of the body is imaged large or small, it can be dealt with by image processing.

Next, the deformation of the shape of the important area 82 will be described with reference to FIGS.

The deformation of the shape of the emphasis area 82 may be caused, for example, when the object to be determined or the characteristic portion (the face portion of the operator) of the object to be determined is imaged without directly facing the imaging unit (camera) 20. The characteristic part of the body or the body to be determined is the imaging unit (camera)
The object is to correct this by image processing when the image is not taken at a predetermined angle with respect to 20, and to cope with it.

For example, it is assumed that the operator's face is imaged slightly sideways with respect to the imaging unit (camera) 20. At this time, as shown in FIG.
When the emphasis area 821 indicated by a broken line in the effective image area (captured image plane) 80 due to 0 is rotated by a predetermined angle around the vertical axis C, the emphasis area reaches a position 822 indicated by a solid line. As seen from the direction perpendicular to the effective image area (captured image plane) 80, the important area 822 rotated about the vertical axis C in this manner is reduced in the left-right direction as shown in FIG. In this way, the rotated focus area 822 is mapped onto the original imaging screen 80 in the mapping area 82.
3 is a new important area.

As shown in FIG. 24 (a), the image processing for deforming the important area 82 in this manner is performed as shown in FIG. 24 (a). According to the angle of the (face portion), the shape of the emphasis area 82 is first deformed to cope with it, and thereafter the effective image area 80 is divided into blocks 84,.
.., And block characteristic values are obtained for each of the blocks 84.

By rotating the emphasis area 82 around an axis perpendicular to the effective image area (captured image plane) 80, the operator tilts his / her head obliquely as shown in FIG. At times, the emphasis area 82 can be made to correspond, and FIG.
As shown in (b) and (c), it is possible to create image data in which the characteristic portion (the operator's face portion) of the object to be determined is just included in the emphasis area 82.

Alternatively, as shown in FIG. 26, if the emphasis area 82 is rotated around the axis C in the oblique direction and the mapping area is transformed into a new emphasis area 82,
As shown in FIGS. 27 (b) and 27 (c), when the operator tilts his / her head slightly downward and downward as shown in FIG. It is possible to create image data that just includes a characteristic portion (the face portion of the operator).

If the shape of the important area 82 is deformed as described above, the angle of the object to be determined or the characteristic portion of the object to be determined with respect to the imaging unit (camera) 20 is not always constant when the image is captured. Even if the characteristic portion of the object to be determined is deformed and imaged, it can be dealt with by image processing.

Next, still another embodiment will be described. In the above embodiment, the current image data and the registered image data are converted into the respective blocks 84.
85, the block characteristic values are compared and collated for each of the current image data and the registered dictionary 421.
In the collation with the registered image data recorded in the memory (or the collation of the current image data with the image data recorded in the continuous collation work memory 411), both image data are represented by one representative value and compared. Thus, the collation processing is simplified and the processing time is shortened.

The representative value representing the image data may be a value uniquely calculated from a set of a plurality of block characteristic values constituting the image data. Specifically, for example, the average value of all the block characteristic values is used. Values can be given.

In this way, the block characteristic values of the blocks 84 having a small area in the emphasis area 82 are treated equally with the block characteristic values of the blocks 85 having a large area in the other areas. Has a great influence on the representative value representing the value of the key region, and as a result, a representative value in which the priority area 82 is emphasized is obtained. And
Since the matching process between the current image data and the registered image data is a process of only comparing a pair of representative values, the matching process becomes extremely easy, and the processing time can be reduced.

Even when the matching process is performed by representing the image data with one representative value as described above, the brightness correction for matching the brightness between the current image and the registered image may be performed.

The registration dictionary 421 has all blocks 84
, 85... Are recorded, and a representative value representing the entire image is obtained at the time of the collation processing, or a previously obtained representative value is recorded in the registration dictionary 421 as registered image data. Is also good.

As described above, the present invention has been described with reference to the embodiment. However, the operation permission / non-permission determining apparatus according to the present invention is not limited to the above-described embodiment, and may be configured as follows.

(1) In the above embodiment, substantially the entire area of an image obtained by imaging is divided into blocks as an effective image area 80 to obtain predetermined image data. The image processing may be performed by setting a predetermined portion, for example, only the central portion, as the effective image region 80 and further setting an important region 82 in the effective image region 80.

(2) In the above embodiment, the markers 81 indicating the four corner positions of the important area 82 are used to display the important area 82 on the monitor 74.
Are displayed as a frame line surrounding the important area 82,
Various modes can be adopted as long as the operator can confirm the important area 82 on the monitor 74, such as a mark indicating the center position of the important area 82.

(3) In the above embodiment, only the one-stage emphasis area 82 divided into blocks having the same area is provided in the effective image area 80 of the captured image.
As shown in FIG. 9, a plurality of stages of important regions 831, 833, and 833 divided into blocks having different areas are provided, and the region 833 to be particularly emphasized is divided into blocks having smaller areas. Is also good.

(4) In the above embodiment, the preliminary collation processing is performed only at the time of the initial collation processing and the return waiting processing. However, the preliminary collation processing may be performed in the continuous collation processing.

(5) In the above embodiment, the camera device 1 which is the main body of the judging device including the image pickup section 20, the registration dictionary 421, the memory 411 for continuous collation work, the image calculating section 30 as collation judging means, and the operation target Although the personal computer 5 is connected to the personal computer 5 by the communication means 6, the personal computer 5 may be provided with the registration dictionary 421, the image calculation unit 30, and the like.

(6) In the above embodiment, the present invention is applied to the determination of the log-in to the personal computer. However, the operation permission / non-permission determining device according to the present invention determines whether or not the opening / closing of the safe is permitted, and the door of the automobile. The present invention can be applied to various determination devices such as a determination device for determining whether to permit opening and closing and starting the engine.

(7) In the above embodiment, the object to be determined is an image of the operator's face captured by the imaging unit 20. However, for example, a doll, a stuffed toy, a key, an ID card, a fingerprint, or the like of the operator's favorite is determined. It can also be a body.

[0190]

As described above, according to the operation permission / non-permission determining apparatus of the present invention, a predetermined area is specified as an important area in an effective image area of an image picked up by an image pickup unit, and the inside of this important area is specified as another area. In order to divide the area into blocks having smaller areas, image data in which the image information in the specific area is more important than the image information in other areas is created. Therefore, if, for example, the area occupied by the object to be determined or the area occupied by the characteristic portion of the object to be determined is regarded as the important area, the difference between the current image data and the image data of the registered dictionary in these parts is sensitively detected. In addition, it is possible to realize more accurate and accurate collation, and to allow a slight difference between the two image data due to a change in an imaging environment such as a background portion, thereby preventing erroneous determination.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a schematic configuration of a camera device that is a main body of a determination device among operation permission / denial determination devices according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating a schematic configuration of a personal computer which is an operation target in the operation permission / non-permission determining apparatus according to the embodiment of the present invention.

FIG. 3 is a part of a flowchart showing a basic flow.

FIG. 4 is the remaining part of the flowchart showing the basic flow.

FIG. 5 is a flowchart showing an initial setting flow.

FIG. 6 is a flowchart showing an initial value flow.

FIG. 7 is a part of a flowchart showing a registration dictionary creation flow.

FIG. 8 is the remaining part of the flowchart showing the registration dictionary creation flow.

FIG. 9 is a flowchart illustrating an imaging processing flow.

FIG. 10 is a flowchart illustrating an image processing flow.

FIG. 11 is a part of a flowchart showing an initial collation processing flow.

FIG. 12 is the remaining part of the flowchart showing the initial collation processing flow.

FIG. 13 is a part of a flowchart showing a continuous matching processing flow.

FIG. 14 is the remaining part of the flowchart showing the continuous collation processing flow.

FIG. 15 is a part of a flowchart showing a flow of a return waiting process;

FIG. 16 is the remaining part of the flowchart showing the return wait processing flow.

FIG. 17 is a data configuration diagram of a registered dictionary.

FIG. 18 is a data configuration diagram of a continuous collation work memory.

FIG. 19 is an explanatory diagram illustrating an example of a captured image displayed on a monitor.

FIG. 20 is an explanatory diagram schematically showing an example of an image processing procedure.

FIG. 21 is an explanatory diagram schematically showing an example of an image processing procedure for moving an important area.

FIG. 22 is an explanatory diagram schematically showing an example of an image processing procedure for enlarging an important area.

FIG. 23 is an explanatory diagram of deformation of an important area.

FIG. 24 is an explanatory diagram schematically showing an example of an image processing procedure for deforming an important area.

FIG. 25 is an explanatory diagram schematically showing an example of an image processing procedure for deforming an important area.

FIG. 26 is an explanatory diagram of deformation of an important area.

FIG. 27 is an explanatory diagram schematically showing an example of an image processing procedure for deforming an important area.

FIG. 28 is an explanatory diagram showing an example of blocks used for a preliminary matching process.

FIG. 29 is an explanatory diagram showing another example of block division of an effective image area.

[Explanation of symbols]

 Reference Signs List 1 camera device (judgment device main body) 10 signal control unit 20 imaging unit 30 image calculation unit (image processing unit, image comparison processing unit) 41 working memory unit 411 continuous matching work memory 42 data holding memory unit 421 registration dictionary 422 ID data Memory 5 Personal computer (operation target) 50 Signal control unit 52 Data holding memory unit 521 ID data memory 6 Communication cable (communication means) 78 CD-ROM reader 80 Effective image area 81 Marker 82 Important area 84 Block in important area 85 Others Block of area

 ──────────────────────────────────────────────────続 き Continued on the front page F-term (reference) EA12 EA16 FA19 FA32 GA19 HA08 JA11 JA13 KA15 9A001 BB03 DD11 HH21 HH23 HH24 JJ01 JJ71 KK54 LL03

Claims (8)

[Claims] An image pickup unit for picking up an image of an object to be determined, image processing means for performing image processing on an image picked up by the image pickup unit to create image data, and image data of the object to be determined are recorded in advance. The registered dictionary is compared with current image data captured by the imaging unit and created by the image processing unit with image data recorded in the registered dictionary. An operation permission / non-permission determining device comprising: a collation determining unit that determines whether the effective image area of the captured image is divided into a plurality of blocks. A block characteristic value representing each block is obtained, and a set of the block characteristic values is created as image data. And specifying a predetermined region as a focus area, actuating permission determining apparatus characterized by dividing the block of the focus area in a small area than blocks of another region. 2. The method according to claim 1, wherein the collation determination unit is configured to compare and collate block characteristic values of a plurality of blocks included in current image data and recorded image data for each corresponding block. The operation permission / non-permission judging device according to claim 1. 3. A comparison and collation for each block corresponding to a block characteristic value of a part of a plurality of blocks included in current image data and recorded image data before collation by the collation determination means. 3. The operation permission / rejection according to claim 2, wherein the collation by the collation determining means is performed only when the preliminary collation result satisfies a predetermined preliminary collation determination criterion. Judgment device. 4. The image processing means performs a division process of dividing an effective image area of a captured image into a plurality of blocks, obtains a block characteristic value representative of each block for each block, and further includes: And calculating a representative value representing the entire captured image as image data from a set of block characteristic values of the image data, and the comparing and judging means compares and compares the representative value between the current image data and the recorded image data. The operation permission / non-permission determining device according to claim 1, wherein the device is configured as follows. 5. The apparatus according to claim 1, wherein in the division processing, the position of the emphasis area is movable. 6. The operation permission / non-permission determination device according to claim 1, wherein in the division processing, the size of the important region can be enlarged or reduced. 7. The image forming apparatus according to claim 1, wherein in the division processing, the emphasis area can be transformed into a mapping area on the captured image plane of an area rotated by a predetermined angle around at least one axis with respect to the captured image plane. Claims 1 to 6
The operation permission / refusal determination device according to any one of the above. 8. The image processing apparatus according to claim 1, further comprising a monitor for displaying an image captured by the image capturing unit, wherein the important area is displayed in the captured image displayed on the monitor. Operation permission / non-permission judgment device.