JP2008052549A - Image processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the efficiency of authentication processing in an image processing system for performing authentication on the basis of an image to be authenticated present in a prescribed area. <P>SOLUTION: A wireless terminal device 6 owned by an authentication target 21 wirelessly transmits identification information and a wireless communication means 2 receives the identification information. An image acquisition means 3 acquires an image of the authentication target and a feature quantity acquisition means 4 acquires a feature quantity. A storage means 4 stores feature quantities and feature quantity determination information while in association with the identification information data of a plurality of targets. When the feature quantity determination information for regulating a determination method different from a standard determination method in association with the received identification information is stored, an authentication means 4 collates the feature quantity stored in association with the identification information with the acquired feature quantity by using the feature quantity determination information to authenticate the authentication target. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing system, and more particularly to an image processing system that improves the efficiency of processing for authenticating an object based on an image.

For example, conventionally, images (images) are digitized, and the presence, type, movement, etc. of objects appearing in the images based on the brightness and color information of each pixel of the images, their time-series changes, spatial distribution, etc. A technique called image recognition for identifying the image has been studied.
With recent improvements in CPU (Central Processing Unit) performance, things that can be identified with this technology have expanded. In recent years, object identification such as humans, dogs, cats, cars, human face identification, and behavior identification Is becoming possible.

As an application to which such image recognition technology is applied, there is a face authentication entrance / exit management system.
When a person enters or exits an area having advanced assets, for example, a room that manages customer information in a company, the face of the target person is photographed by a camera, and the face is recognized by image recognition technology from the photographed image. This is a system for ensuring security by extracting the feature amount, comparing it with a previously extracted facial feature amount database, and unlocking and locking the door based on the match or mismatch.

FIG. 5 shows a configuration example of the face authentication entrance / exit management system.
The face authentication entrance / exit management system of this example includes a trigger device 111, a camera 112, a face authentication device 113, an electric lock 114, and transmission media 121, 122, and 123.
FIG. 5 shows a door 115 for managing entry / exit and a person 131 to be authenticated.
The camera 112 has an imaging unit and a video data output unit, images the periphery of the face of the authentication target 131, and transmits the captured video to the face authentication device 113 via the transmission medium 122.
The trigger device 111 has a trigger signal output unit, and gives a face authentication start trigger to the face authentication device 113 via the transmission medium 121.

The face authentication device 113 includes a video data input unit, a feature amount extraction processing unit, a feature amount verification unit, a feature amount database unit, an electric lock control signal output unit, and a trigger signal input unit.
In response to the trigger signal input, the face authentication device 113 performs feature amount extraction processing, in this example, face feature amount extraction processing, on the input video data by the feature amount extraction processing unit. Further, the extracted feature quantity is collated with the information acquired from the feature quantity database unit by the feature quantity collation unit, the presence / absence of a matching face is determined, and if it exists, the entry / exit permission information of the person is extracted, Decide whether to enter or leave. When it is determined that entry / exit is possible, a signal is output from the electric lock control signal output unit and transmitted to the electric lock 114 via the transmission medium 123.
The electric lock 114 has an electric lock control signal input unit, and opens and closes the lock according to the input signal.

JP 2006-101028 A

However, in the face authentication entrance / exit management system as described above, the extracted feature quantity is compared with the feature quantity database in the face authentication apparatus 113 until a matching face is found from the first person information registered in the database. Since the processes are performed sequentially, there is a problem that the authentication process takes a long time.
In particular, the larger the parameter registered in the feature amount database, the greater the load of collation processing in the face authentication device 113, and the longer it takes to output the collation result. If no matching face is found, verification with all registered data is necessary, and the load is maximized.

  Further, in the face authentication entrance / exit management system as described above, the greater the number of registered people in the feature database, the greater the probability that a person with a similar feature exists in the database, and the similar feature As the number of persons having increased, the number of false authentications increased and the accuracy rate in authentication decreased.

  SUMMARY An advantage of some aspects of the invention is that it provides an image processing system capable of improving the efficiency of processing for authenticating an object based on an image.

In order to achieve the above object, in the image processing system according to the present invention, authentication is performed on the basis of an authentication target image existing in a predetermined area with the following configuration.
That is, the wireless terminal device has a function of wirelessly transmitting identification information for identifying the authentication target, and is possessed by the authentication target.
The wireless communication means receives the identification information transmitted wirelessly from the wireless terminal device possessed by the authentication target existing in the predetermined area.
An image acquisition unit acquires an authentication target image existing in the predetermined area.
A feature amount acquisition unit acquires a feature amount related to the authentication target based on the image acquired by the image acquisition unit.
The storage unit stores feature amounts in association with identification information for a plurality of targets that can be authentication targets, and stores feature amount determination information that defines how to determine feature amounts for one or more targets. To do.
When the authentication unit stores feature quantity determination information that defines a determination method different from a standard determination method in association with the identification information received by the wireless communication unit, Using the determination method defined by the feature amount determination information, the feature amount stored in the storage unit in association with the identification information and the feature amount acquired by the feature amount acquisition unit are collated, Authenticates the authentication target. In other cases, the authentication unit uses the standard determination method and associates the feature amount stored in the storage unit with the identification information and the feature amount acquired by the feature amount acquisition unit. And authenticate the authentication target.

Therefore, when performing authentication for the authentication target based on the authentication target image existing in the predetermined area, the authentication target identification information existing in the predetermined area is acquired and stored in correspondence with the identification information. For example, the feature quantity corresponding to the identification information of the authentication target that does not exist in the predetermined area is not referred to at the time of collation. Processing time and processing load required for processing can be reduced, and erroneous recognition at the time of collation can be reduced.
In addition, since it is possible to specify the method of determining the feature amount for each authentication target, for example, a determination method suitable for the feature for each authentication target can be used, and thus, the verification process is required. Processing time and processing load can be reduced, and accuracy of collation can be improved.
In this way, the process of authenticating the target based on the image can be made efficient.

Here, various areas may be used as the predetermined area. For example, an area in front of a room or an area in front of a gate can be used.
In addition, various objects may be used as the authentication target. For example, a person or a thing may be used. In the case of a person, the user wears a wireless terminal device or the like, and in the case of an object, the wireless terminal device is attached.
Various kinds of information may be used as identification information for identifying an authentication target, and for example, a personal number may be used.
Further, the means for acquiring the authentication target image existing in the predetermined area can be configured using, for example, a camera that captures the predetermined area.

Various features may be used for the authentication target, and various numbers may be used. For example, when the authentication target is a person, a value (quantity) representing the feature of the face can be used, and when the authentication target is an object, the features such as the shape and pattern on the surface of the object can be used. The value (amount) represented can be used.
In addition, as an aspect for acquiring a feature amount related to an authentication target based on an image, for example, an aspect is used in which an image is processed and a feature amount used for collation is calculated by calculation for an object reflected in the image. Can do.

In addition, as a plurality of targets for storing feature amounts in advance for reference at the time of collation, various items that can be authentication targets may be used, and various numbers may be used. For example, when managing entry into a predetermined room, feature quantities can be measured and stored in advance for a plurality of persons permitted to enter the room.
Various standard determination methods for feature quantities may be used. For example, a standard one suitable for many authentication targets may be used.
Moreover, various aspects may be used as the object (identification information) and the number of features (identification information) for storing the feature amount determination information that defines a determination method different from the standard determination method for the feature amount.

Further, various determination methods different from the standard determination method for feature amounts may be used. For example, the order (priority order) of determining a plurality of feature amounts one by one is used. How to make a special decision, how to make a special calculation formula for calculating a predetermined feature, and how to make a special weight for each degree of matching of multiple features Or a determination method in which the type or number of one or more feature quantities used for matching among a plurality of feature quantities is specified, or a combination of two or more of these.
Moreover, as a memory | storage means, it can comprise using a memory, for example.

  In addition, various modes may be used as modes for performing verification by comparing feature quantities. For example, the degree of coincidence (for example, closeness of values) between an acquired feature quantity and a referenced feature quantity Can be used when it satisfies the preset conditions, it can be regarded as matching and it is determined that the authentication has been performed properly. A mode in which it is determined that the device has been properly authenticated, or a mode in which it is determined that the device has been properly authenticated when one or more predetermined feature quantities match can be used.

  In addition, for example, when the feature quantity to be referred to is one value, an aspect that determines that the acquired feature quantity matches the one value when the acquired feature quantity matches the one value, or the acquired feature quantity Can be used when both match with one value within a predetermined tolerance, and when the feature value to be referred to is a range of values having a width. Can use a mode in which when the acquired feature amount is within the range, it is determined that the two match.

In addition, when a plurality of authentication targets exist in a predetermined area and a plurality of pieces of identification information are acquired, for example, a mode in which the plurality of pieces of identification information are collated one by one in order can be used. . In addition, when the final authentication result is obtained in the middle of collating one by one for a plurality of pieces of identification information, the subsequent collation need not be performed.
Further, when a plurality of authentication targets are displayed in an image of a predetermined area, for example, a mode in which only one of the plurality of authentication targets is collated, or the plurality of authentication targets is used. It is possible to use a mode in which collation is performed one by one in order.

  In addition, as a wireless communicable range of a wireless communication unit that receives identification information wirelessly transmitted from a wireless terminal device and an image acquisition range (for example, an imaging range of a camera) of an image acquisition unit that acquires an image to be authenticated, The ranges may not necessarily coincide with each other, and various ranges may be used as long as they are practically effective. For example, a range covered in common by both is a predetermined region.

The present invention can also be provided as a method, a program, a recording medium, and the like.
In the method according to the present invention, each unit executes various processes in the apparatus or system.
The program according to the present invention is executed by a computer constituting the apparatus or system, and various functions are realized by the computer.
In the recording medium according to the present invention, a program to be executed by a computer constituting the apparatus or system is recorded so as to be readable by the input means of the computer, and the program causes the computer to execute various processes.

  As described above, according to the present invention, when authentication is performed on an authentication target based on an authentication target image existing in a predetermined area, a method for determining a feature amount is defined for each authentication target. Therefore, for example, it is possible to use a determination method suitable for the feature of each authentication target, thereby reducing the processing time and processing load required for the matching process, and improving the matching accuracy. Can do. In this way, the process of authenticating the target based on the image can be made efficient.

Embodiments according to the present invention will be described with reference to the drawings.
FIG. 1 shows a configuration example of a face authentication entrance / exit management system according to an embodiment of the present invention.
The face authentication entrance / exit management system of this example includes a trigger device 1, a wireless base station device 2, a camera 3, a face authentication device 4, an electric lock 5, a wireless tag 6, a trigger device 1, and a face authentication device. 4, a transmission medium 12 connecting the radio base station apparatus 2 and the face authentication apparatus 4, a transmission medium 13 connecting the camera 3 and the face authentication apparatus 4, an electric lock 5 and the face authentication apparatus 4. The transmission medium 14 is connected.
Further, FIG. 1 shows a door 7 where entry / exit is managed, a person 21 to be authenticated, and a wireless area 22 in which wireless communication can be performed by the wireless base station device 2.
The trigger device 1, the wireless base station device 2, the camera 3, and the electric lock 5 are provided on the door 7, the wireless tag 6 is attached to the person 21, and the face authentication device 4 is, for example, the door 7. It is provided in an arbitrary place such as near or separated.

In this example, the authentication target is a person 21 who wants to move to the area beyond the door 7 and possesses the wireless tag 6.
The door 7 has a door knob, for example, and is a door that is manually opened and closed. As other configuration examples, a sliding door, an automatic door, a gate such as a railway ticket gate, a gate whose entrance and exit is controlled by a bar such as a crossing may be used.
Further, the wireless area 22 is a range in which wireless radio waves from the wireless base station device 2 can reach (wireless communication with the wireless base station device 2 is possible).

FIG. 2 shows a detailed configuration example of each of the component devices 1 to 6 provided in the face authentication entrance / exit management system of this example and their connection relationship.
The camera 3 is a device having an imaging unit 51 and a video data output unit 52, and is, for example, an analog type camera. As shown in FIG. 1, the camera 3 is installed at a position where the face of the person 21 to be authenticated can be photographed.
The imaging unit 51 converts the input optical signal (video) into an electrical signal.
The video data output unit 52 externally outputs the converted electrical signal (video data) and transmits it to the face authentication device 4 via the transmission medium 13 in this example.
Note that a network camera or the like may be used as another configuration example. When a network camera is used, the camera 3 includes, for example, an image codec that digitizes the captured video and performs compression processing (if necessary) in addition to the configuration of this example.

The wireless tag 6 is a device having a memory unit 81 and an antenna unit 82, and is, for example, a passive RF-ID tag. As another configuration example, an active RF-ID tag, a ZigBee sensor node, or the like may be used.
In the memory unit 81, information for identifying an individual, such as an employee number, is recorded in advance in a nonvolatile manner. In this example, a personal number is used as information for identifying an individual. As another configuration example, information other than a number such as an alphabet may be used as information for identifying an individual.
The antenna unit 82 transmits and receives data to and from the radio base station apparatus 2 by radio. In this example, in particular, data of personal identification information (in this example, personal number) stored in the memory unit 81 is transmitted.

The radio base station apparatus 2 is an apparatus having an antenna unit 41, a tag data output unit 42, and a control unit 43. The radio base station apparatus 2 is installed so that the radio area 22 includes an area in front of the door 7. The transmission medium 12 shown in FIG. 1 includes transmission media 12a and 12b shown in FIG.
The antenna unit 41 transmits and receives data to and from the wireless tag 6 wirelessly.
The tag data output unit 42 outputs the data received from the wireless tag 6 to the outside, and in this example, transmits it to the face authentication device 4 via the transmission medium 12b.
The control unit 43 receives and executes a control command from the outside, and in this example, inputs the control command from the face authentication device 4 via the transmission medium 12a.

The face authentication device 4 includes a video data input unit 61, a feature amount extraction processing unit 62, a feature amount database unit 63, a feature amount collation unit 64, a filtering unit 65, a tag data management unit 66, and tag data input. This is a device having a unit 67, an electric lock control signal output unit 68, and a trigger signal input unit 69, and is composed of, for example, a computer.
The video data input unit 61 receives an externally input video and converts it into a signal that can be handled inside the apparatus. In this example, a video (image) is received from the camera 3 via the transmission medium 13.

The feature amount extraction processing unit 62 performs a process of extracting a feature amount from a video using an image recognition algorithm, in this example, a process of extracting a face feature amount.
The feature quantity database unit 63 holds feature quantities of all persons to be authenticated, holds permission / rejection permission information for these persons, and includes feature quantity determination information indicating how to determine the feature quantities for these persons. Hold. Here, the feature amount of each person is calculated in advance, for example. In this example, the feature amount database is configured by the information held in the feature amount database unit 63.
In addition, various things may be used as a feature amount of a person's face, for example, a feature amount related to eyes, a width between eyes, a nose, an arrangement of eyes and nose, and the like can be used. Also, feature quantities relating to distance, shape, brightness, hue, and the like can be used.

FIG. 3 shows an example of the feature amount database.
The feature amount database 91 includes a personal number sequence 101, a feature amount sequence 102 related to the first feature amount A, a feature amount sequence 103 related to the second feature amount B, and a feature amount sequence 104 related to the third feature amount C. , An entry / exit permission information sequence 105 and a feature amount determination information sequence 106.
The personal number column 101 stores the personal number of each person.
Each feature amount column 102, 103, 104 stores a feature amount related to each person (for example, a numerical value representing a feature). In this example, three types of feature quantities are stored for each person, but an arbitrary number of one or more kinds of feature quantities may be stored.

  The entry / exit permission information column 105 stores information indicating “permitted” or “non-permitted” as information indicating permission to enter or leave the room with the door 7 as an entrance. In this example, since there is only one area for entry / exit management, only one entry / exit permission information column 105 is provided. However, as another configuration example, entry / exit management for a plurality of areas is performed. When performing, the entrance / exit permission information column 105 may be provided for each area, and a plurality of columns may be provided as a whole. In this example, a value having a range is used as the feature amount. However, as another configuration example, a single value having no range may be used.

The feature amount determination information column 106 stores feature amount determination information indicating how to determine the feature amount for each person.
In this example, first, a standard determination method is set as a reference for how to determine the feature amount. In the standard determination method, for example, the order of determination is set in advance, such as the order of the feature amount A, the feature amount B, and the feature amount C, and the standard calculation for calculating each feature amount is performed. Expressions (arithmetic expressions) are set, and there is no weighting for each feature amount.

In the feature amount determination information, a determination method different from the standard determination method is specified, for example, it is specified that a priority order for determining a plurality of feature amounts in an order different from the standard determination method is used, Alternatively, it is specified that a calculation formula different from the standard calculation formula is used for a predetermined feature quantity, or that a weight different from the standard judgment method is given for a plurality of feature quantities. Or, a combination of two or more of these can be defined.
Specifically, for each person, a determination method suitable for authenticating each person can be set. For example, since a certain person has a large feature in the width of the eye, a feature of the width of the eye Since the determination of the amount is the highest priority, and for some people, the formula for calculating the arrangement of the eyes and nose is different from the others, it is highly compatible and the formula is It is possible to use or weight these feature quantities based on the importance of various feature quantities for a certain person.

In the example of FIG. 3, information “priority order A” or “priority order B” that designates use of a predetermined priority order, or “calculation formula A” or “calculation formula” that designates use of a predetermined calculation formula. Information “B” is set. In addition to these, various feature amount determination information may be used.
In this example, it is assumed that the standard determination method is set to be used when no information is set in the feature amount determination information sequence 106.

The feature amount matching unit 64 compares the feature amount related to the face of the person 21 extracted from the video with the feature amount of the referenced feature amount database 91, identifies the photographed person, and determines whether or not the authentication target is allowed to enter or leave. To decide.
The filtering unit 65 narrows down the reference feature amount candidates to be collated in the feature amount collating unit 64 according to the personal number list acquired based on the data from the wireless tag 6. In this example, a visitor database is constructed from a list of reference feature values obtained by narrowing down.

The tag data input unit 67 receives external tag data input and converts it into a signal that can be handled inside the apparatus. In this example, tag data is received from the radio base station apparatus 2 via the transmission medium 12b. In particular, the tag data includes personal identification information (in this example, a personal number) stored in the wireless tag 6.
The tag data management unit 66 manages the tag data input via the tag data input unit 67 by storing it in a memory, for example, and sends a control command to the radio base station apparatus 2 via the transmission medium 12a. In this example, a list of personal numbers acquired from the wireless tag 6 present in the wireless reachable range (wireless area 22) of the wireless base station device 2 is currently held.
The electric lock control signal output unit 68 outputs a lock open / close signal to the outside, and transmits an open / close signal to the electric lock 5 via the transmission medium 14 in this example.
The trigger signal input unit 69 receives an externally input trigger signal, and in this example, receives the trigger signal from the trigger device 1 via the transmission medium 11.

The electric lock 5 is a device having an electric lock control signal input unit 71 and a lock opening / closing unit 72.
The electric lock control signal input unit 71 receives an externally opened / closed lock signal, and in this example, receives the open / close signal from the face authentication device 4 via the transmission medium 14.
The lock opening / closing unit 72 opens / closes the lock according to the input signal, and in this example, opens / closes the lock according to the received open / close signal. In this example, the configuration for controlling the opening and closing of the lock by a combination with a door that is manually opened and closed is shown. However, as another configuration example, a device for controlling the opening and closing of an automatic door by a combination with an automatic door is shown. The provided structure can also be used.

The trigger device 1 is a device having a switch unit 31 and a trigger signal output unit 32, and is composed of a mat with a switch, for example. As another configuration example, the trigger device 1 may be configured using an infrared sensor device, a button chime, an IC card reader device, or the like.
The switch unit 31 switches the switch from off to on when the authentication object comes to the authentication place. In this example, the person 21 is placed on the trigger device 1 placed in front of the door 7. In response to the switch.
The trigger signal output unit 32 externally outputs a trigger signal in response to the switch of the switch unit 31 being turned on, and transmits the trigger signal to the trigger signal input unit 69 via the transmission medium 11 in this example.

  The transmission medium 13 connects the video data output unit 52 of the camera 3 and the video data input unit 61 of the face authentication device 4, and for example, a coaxial cable that transmits analog video can be used. As another configuration example, a network cable or a serial cable that transmits digital video may be used.

The transmission medium 12a connects the control unit 43 of the radio base station apparatus 2 and the tag data management unit 66 of the face authentication apparatus 4, and for example, a network cable can be used. As another configuration example, a contact cable that transmits a contact signal, a serial cable, or the like may be used.
The transmission medium 12b connects the tag data output unit 42 of the radio base station apparatus 2 and the tag data input unit 67 of the face authentication apparatus 4, and for example, a network cable or a serial cable can be used.
In this example, the transmission medium 12a and the transmission medium 12b are configured as separate media. However, as another configuration example, a single network cable may be used.

The transmission medium 14 connects the electric lock control signal input unit 71 of the electric lock 5 and the electric lock control signal output unit 68 of the face authentication device 4, and a network cable can be used, for example. As another configuration example, a contact cable or a serial cable may be used.
The transmission medium 11 connects the trigger signal output unit 32 of the trigger device 1 and the trigger signal input unit 69 of the face authentication device 4. For example, a network cable can be used. As another configuration example, a contact cable that transmits a contact signal may be used.
In this example, each of the transmission media 11, 12a, 12b, 13, and 14 may include a network device such as an amplifier, a router, a switch, and a hub, a video distributor, and the like, and a wireless LAN ( Local Area Network) or the like may be used.

  Here, in this example, in order to clarify each function and make the explanation easy to understand, the wireless base station device 2, the camera 3, the face authentication device 4, and the electric lock 5 are configured as separate devices that are physically separated from each other. However, as another configuration example, devices integrated in any combination may be used. Specific examples include a device in which the camera 3 and the face authentication device 4 are integrated, a device in which the camera 3 and the wireless base station device 2 are integrated, and a device in which all these devices are integrated. Also good.

In this example, for simplicity of explanation, the camera 3, the wireless tag 6, the wireless base station device 2, the face authentication device 4, the electric lock 5, and the trigger device 1 are shown as a single unit. It is also possible to configure a plurality of units as appropriate.
Further, in this example, the door 7 is shown as an example of the management target by the entrance / exit management system. However, for example, when the system requirements are such that entrance management is performed but entry prevention is not required, the door is used. 7 and the electric lock 5 may not be provided.
Furthermore, if it is a system requirement that you do not need to stop admission but want to be careful about admission, for example, you can use the electric lock 5 for voice warnings or warnings for admission when disapproved. It can be replaced with a light alarm device. Further, in this case, a warning unit by voice or light is provided on the wireless tag 6 side, and the wireless tag 6 issues a warning in response to a warning signal from the wireless base station device 2 when it is not permitted. Is also possible.

Next, an example of the flow of processing at the time of entrance / exit performed in the face authentication entrance / exit management system of this example is shown.
First, the person 21 to be authenticated stands in front of the door 7 while holding the wireless tag 6. In front of the door 7, for example, a mat with a switch is installed as the trigger device 1, and the person 21 gives the trigger device 1 an action, for example, an act of stepping on the mat.
The trigger device 1 detects an action given by the person 21 by the switch unit 31 and outputs a trigger signal from the trigger signal output unit 32. The output trigger signal is input to the trigger signal input unit 69 of the face authentication device 4 via the transmission medium 11.
When the face authentication device 4 detects the trigger signal input to the trigger signal input unit 69, the face authentication device 4 outputs a tag confirmation command signal from the tag data management unit 66. In addition, the video data input unit 61 starts video reception processing.

Here, as a process in which the camera 3 converts the optical signal into an electrical signal by the imaging unit 51 and outputs the electrical signal from the video data output unit 52, a configuration that is always performed may be used, or the face authentication device 4 may be used. A configuration that sequentially outputs in response to a request from may be used. For example, when the camera 3 is an analog camera, it is often considered that the configuration is always output, and when the camera 3 is a network camera, it is often assumed that the configuration is sequential output. .
When the output of the camera 3 is a sequential output, the face authentication device 4 transmits a video request from the video data input unit 61 to the camera 3.
The video data output from the video data output unit 52 of the camera 3 is input to the video data input unit 61 of the face authentication device 4 via the transmission medium 13.

On the other hand, the tag confirmation command signal output from the tag data management unit 66 of the face authentication apparatus 4 is input to the control unit 43 of the radio base station apparatus 2 via the transmission medium 12a.
When the radio base station apparatus 2 detects the tag confirmation command signal input to the control unit 43, the radio base station apparatus 2 wirelessly transmits a personal number request from the antenna unit 41 to the radio area 22.
When receiving the personal number request from the antenna unit 82, the wireless tag 6 existing in the wireless area 22 reads the personal number stored in the memory unit 81 and wirelessly transmits the personal number from the antenna unit 82. Here, this processing is performed not only by the wireless tag 6 possessed by the person 21 to be authenticated but also by all wireless tags existing in the wireless area 22.
In this example, the wireless tag 6 side passively transmits the personal number in response to a request from the wireless base station device 2. However, as another configuration example, the wireless tag 6 side periodically transmits the personal number. A configuration for performing spontaneous transmission may be used.

The radio base station apparatus 2 receives the personal number transmitted from the radio tag 6 by the antenna unit 41 and outputs tag data including the personal number from the tag data output unit 42. The output personal number is input to the tag data input unit 67 of the face authentication device 4 via the transmission medium 12b.
The face authentication device 4 receives the personal number input to the tag data input unit 67 and passes it to the tag data management unit 66. Further, the tag data management unit 66 creates a list of personal numbers input after the output of the tag confirmation command signal, that is, a list of personal numbers stored in the wireless tag 6 present in the wireless area 22; The data is passed to the filtering unit 65.
The face authentication device 4 uses the filtering unit 65 to extract from the feature amount database unit 63 only the feature amount, entry / exit permission information, and feature amount determination information for the personal number included in the personal number list, and thereby the resident database. Is generated and passed to the feature amount extraction processing unit 62 and the feature amount matching unit 64.

  FIG. 4 shows an example of a visitor database 92 created from the feature quantity database 91. In this example, among the information on a plurality of personal numbers included in the feature amount database 91, the personal number corresponds to the wireless tag 6 (three wireless tags in this example) that currently exists in the wireless area 22. Only the information about 1, 3, and 4 is extracted by the filtering unit 65, and the visitor database 92 is configured by the extracted information. That is, the visitor database 92 corresponds to a subset database of the feature database 91 obtained by filtering only the users currently existing in the wireless area 22 with respect to the feature database 91.

Further, the face authentication device 4 receives the video data input to the video data input unit 61, converts it into a signal that can be handled inside the device, and passes it to the feature amount extraction processing unit 62. As this conversion, for example, digital conversion is used when the input video data is analog data, and expansion conversion is used when the input video data is digital compression data.
The feature amount extraction processing unit 62 performs a process of extracting the face feature amount from the video using the image recognition algorithm on the received video data, and holds the obtained feature amount in the memory.

In this case, in this example, the feature amount extraction processing unit 62 refers to the feature amount determination information corresponding to each personal number included in the visitor database 92, and calculates the personal number using a standard calculation formula. The feature amount is calculated using the formula, and the feature amount is calculated using the calculation formula for the personal number that is defined to use a calculation formula different from the standard calculation formula.
In addition, when there are two or more calculation formulas corresponding to each personal number included in the visitor database 92, for example, the calculation results of the feature values by all the calculation formulas may be obtained in advance and stored in the memory. However, as an example, when collation is performed in order for each individual number included in the visitor database 92, the feature amount is calculated using the calculation formula designated for each collation when performing each collation. Further, if the configuration is such that the calculation result of the feature value based on the already used calculation formula is stored in the memory and the stored calculation result is used when the same calculation formula is specified later, the efficiency is improved. Is achieved.

The feature amount matching unit 64 checks the feature amount related to the face obtained by the feature amount extraction processing unit 62 with one or more reference feature amounts stored in the visitor database 92, determines whether or not there is a match, and matches If there is a reference feature amount to be entered, the entry / exit permission information of the person having the reference feature amount is extracted.
In this case, in this example, the feature amount matching unit 64 refers to the feature amount determination information corresponding to each individual number included in the visitor database 92, and the personal number for which priority order, weighting, and the like are defined. The collation is performed according to the specified contents.

As a specific example, when collation is performed using each determination method, the individual numbers are sorted in the order of the personal numbers stored in the visitor database 92 (in the example of FIG. 4, in the order of 1, 3, 4). In addition, it is determined whether or not the feature amount related to the face of the person 21 calculated from the image satisfies all feature amounts (in this example, feature amounts A, B, and C) corresponding to the individual numbers. Is determined to be a person corresponding to the personal number and the collation process is terminated. On the other hand, when a part of the personal number is not satisfied, the process proceeds to collation for the next personal number.
Here, when collation is performed using a standard determination method, collation is performed according to a standard determination method defined in advance.

Further, as a determination method different from the standard determination method, for example, when a priority order is defined, the personal numbers are collated in the order of feature amounts according to the priority order. For example, if the first feature value of the priority order is very suitable for the person but not suitable for many other people, only the first feature value is determined, Since it is determined that the number of people is not large, the second and subsequent feature amount determination processing can be omitted for many other people, and the efficiency of matching processing can be improved.
In addition, as a determination method different from the standard determination method, for example, when weighting for a plurality of feature amounts is defined, the degree of coincidence of each feature amount (close to the value of the feature amount) Degree) and multiplying by a weighting coefficient corresponding to each feature amount, and the sum of the multiplication results for a plurality of feature amounts (here, the larger one is the higher the matching degree) is obtained, and the sum When the value is equal to or greater than a predetermined threshold (or exceeds a predetermined threshold), it can be determined that the person corresponds to the personal number.

Further, as a determination method different from the standard determination method, for example, when a calculation formula used for calculating a predetermined feature value is specified, the feature number extraction processing unit 62 specifies the personal number. The feature amount is calculated using the calculated formula.
In addition, the determination method shown here is an illustration, and various other determination methods may be used for collation. As another determination method, for example, it is possible to define the number (one or more) of feature quantities to be collated and the type for each personal number.

When the extracted entry / exit permission information is permission, the feature amount matching unit 64 outputs a signal instructing unlocking from the electric lock control signal output unit 68. The output unlocking signal is input to the electric lock control signal input unit 71 of the electric lock 5 through the transmission medium 14.
When the electric lock 5 detects the unlocking signal input to the electric lock control signal input unit 71, the electric lock 5 is unlocked by the lock opening / closing unit 72.

  As described above, in this example, the face authentication is performed by making the visitor database 92 narrowed down from the feature amount database 91 by the personal number obtained from the wireless tag 6 existing in the wireless area 22 into the extracted feature amount comparison target. The load applied to the device 4 can be reduced, and high-speed collation processing can be performed. For example, even when the number of people registered in the feature amount database 91 increases, the increase in the number of people actually existing in the wireless area 22 at a certain time is usually as large as the increase in the feature amount database 91. In other words, it can be said that this is a method in which the increase in load is small with respect to the increase in the number of registered parameters in the feature amount database 91.

  Further, in this example, by making the visitor database 92 a target for collation of the extracted feature value, it is possible to reduce the probability that a person having a feature value similar to the extracted feature value exists on the database. For example, in FIG. 4, in the extraction process performed for the person corresponding to the personal number 1, when the feature quantity A is calculated to be 215, the feature quantity B is 56, and the feature quantity C is 35, In the method of directly collating with the quantity database 91, the person with the personal number 1 and the personal number 5 falls into the state, and it becomes impossible to determine which is the person. On the other hand, in the method in which the visitor database 92 is a target for comparison, the person with the personal number 5 is not in the wireless area 22 and is therefore excluded in advance, and the person with the personal number 1 is not mistaken. Thus, it can be said that the method of this example is a method of reducing the probability of erroneous authentication with respect to an increase in the registration parameter of the feature amount database 91.

  Further, in this example, it is possible to set a method for determining a feature amount (feature amount determination information) for each person (each individual number), and if it is set, matching processing is performed accordingly. Thus, for example, it is possible to perform matching by effectively using feature portions that are particularly applicable to each person, thereby reducing the processing time and processing load required for the matching process for each person, The accuracy of each verification can be improved.

  Here, in this example, the configuration in which face authentication is performed by using a trigger signal sent from the trigger device 1 to one person 21 standing in front of the door 7 is shown. A camera is installed at a position where the entire area 22 can be imaged, and the number of wireless tags 6 staying in the wireless area 22 is constantly observed. It is also possible to use a configuration in which face authentication processing is performed on all persons staying in the wireless area 22, and in such a configuration, authentication speeding up by making the visitor database 92 a target for verification is extremely high. In addition, it is possible to realize entrance / exit management in a place where a plurality of people pass simultaneously. In addition, the authentication process about a some person may be performed in parallel, for example, or may be performed in order for every person.

  Further, in this example, the configuration for managing entry into the room having the door 7 is shown. However, as another configuration example, when there is no door, when a person enters the room to be managed, the surrounding area is displayed. If there is another person, it may be unclear whether or not he has been authenticated by the system. On the other hand, as an example, the wireless base station apparatus 2 and the wireless tag 6 attached to a person are allowed to perform wireless communication in both directions. First, the wireless tag 6 transmits a personal number to the wireless base station apparatus 2. Next, when this personal number is legitimately authenticated by the face authentication device 4 based on the feature quantity of the person's face image, an authentication confirmation signal indicating that fact is transmitted from the wireless base station device 2 to the wireless tag 6. The wireless tag 6 can be configured to perform a predetermined operation such as sounding, vibrating, or emitting light in response to receiving the authentication confirmation signal. In such a configuration, when a person entering the room or the like is authenticated, the wireless tag 6 attached to the person performs a predetermined operation, so that it is possible to grasp that the person has been authenticated. it can. Thereby, it can prevent misidentifying other person's authentication as own authentication.

  As described above, in the face authentication entrance / exit management system of this example, the video imaging device (camera 3 in this example), the image recognition processing device (face authentication device 4 in this example), the wireless base station device 2, the wireless In an image recognition system having a terminal device (in this example, a wireless tag 6), narrowing down data targeted for feature amount matching according to ID information (in this example, a personal number) sent from the wireless terminal device In addition, it is possible to define a method for determining a feature amount for each authentication target (each ID information) and to perform a matching process according to the method.

  Therefore, in the face authentication entrance / exit management system of this example, it is possible to perform the collation process at high speed by narrowing down the collation target even when, for example, there are many parameters registered in the feature amount database 91. Faster authentication can be realized. Specifically, for example, when a device having a performance equivalent to that of a conventional device is used, simultaneous authentication of a larger number of people can be realized within the same time. Alternatively, instead of increasing the speed of authentication, for example, even when using a cheaper and slower apparatus than the conventional one, the authentication at the same speed as the conventional one can be realized.

  As described above, in this example, it is possible to reduce the load of collation processing in the face authentication apparatus that has been increased in proportion to the number of parameters registered in the feature amount database 91 in the related art. Can reduce the influence of the size of the parameter registered on the face authentication accuracy, and achieve high-precision authentication that is less affected by the size of the parameter registered in the feature database 91 can do.

  In addition, in the face authentication entrance / exit management system of this example, it is possible to define the method of determining the feature amount for each authentication target, for example, using a determination method suitable for each authentication target. As a result, it is possible to increase the speed and load of the authentication process and improve the accuracy of the authentication process.

  In the face authentication entrance / exit management system (an example of an image processing system) of this example, a wireless terminal device is configured by the wireless tag 6 that wirelessly transmits identification information (in this example, a personal number), and the wireless base station The wireless communication means is configured by the function of the station apparatus 2 receiving the identification information wirelessly transmitted from the wireless tag 6, and the image acquisition means is configured by the function of the camera 3 capturing and acquiring an image of a predetermined area. The face authentication device 4 includes a feature amount acquisition unit having a function of acquiring a feature amount related to an authentication target (in this example, a person 21) based on the acquired image. The storage means is configured by the function of storing the feature amount referred to for the identification information and the content of the feature amount determination information (in this example, the feature amount database 91) in the memory, and the face authentication device 4 stores the feature information. The feature amount and the feature amount by matching feature acquired based on the contents of the determination information function by the authentication means for performing authentication is configured.

Here, the configuration of the system and apparatus according to the present invention is not necessarily limited to the configuration described above, and various configurations may be used. The present invention can also be provided as, for example, a method or method for executing the processing according to the present invention, a program for realizing such a method or method, or a recording medium for recording the program. It is also possible to provide various systems and devices.
The application field of the present invention is not necessarily limited to the above-described fields, and the present invention can be applied to various fields.
In addition, as various processes performed in the system and apparatus according to the present invention, for example, the processor executes a control program stored in a ROM (Read Only Memory) in hardware resources including a processor and a memory. A controlled configuration may be used, and for example, each functional unit for executing the processing may be configured as an independent hardware circuit.
The present invention can also be understood as a computer-readable recording medium such as a floppy (registered trademark) disk or a CD (Compact Disc) -ROM storing the control program, and the program (itself). The processing according to the present invention can be performed by inputting the program from the recording medium to the computer and causing the processor to execute the program.

It is a figure which shows the structural example of the face authentication entrance / exit management system which concerns on one Example of this invention. It is a figure which shows the detailed structural example of a face authentication entrance / exit management system. It is a figure which shows an example of the feature-value database. It is a figure which shows an example of a mode that a visitor database is acquired from a feature-value database. It is a figure which shows an example of a face authentication entrance / exit management system.

Explanation of symbols

  1, 111 .. Trigger device, 2 .... Wireless base station device, 3, 112 ... Camera, 4, 113 ... Face authentication device, 5, 114 ... Electric lock, 6 ... Wireless tag, 7, 115 ... · Doors, 11 to 14, 121 to 123 · · Transmission medium, 21, 131 · · Authentication object, 22 · · Wireless area, 31 · · Switch portion, 32 · · Trigger signal output portion, 41 · · Antenna portion, 42 ..Tag data output unit 43 ..Control unit 51 ..Imaging unit 52 ..Video data output unit 61 ..Video data input unit 62 ..Feature extraction processing unit 63. 64 .. Feature amount matching unit 65.. Filtering unit 66.. Tag data management unit 67.. Tag data input unit 68.. Electric lock control signal output unit 69. 71.・ Electric lock control signal input section 72 ・ ・ Lock opening / closing section 81 ・ ・ Memory section 82 ・ ・ Antenna section 91 ・ ・ Feature database 92 ・ ・ Resident database 101 ・ ・ Personal number sequence 102 to 104 .. Feature amount sequence 105.. Entrance / exit permission information sequence 106.. Feature determination information sequence

Claims (1)

In an image processing system that performs authentication based on an authentication target image existing in a predetermined area,
A wireless terminal device possessing a function of wirelessly transmitting identification information for identifying the authentication target and possessed by the authentication target;
Wireless communication means for receiving identification information transmitted wirelessly from a wireless terminal device possessed by an authentication target existing in the predetermined area;
Image acquisition means for acquiring an image to be authenticated existing in the predetermined area;
Feature quantity acquisition means for acquiring a feature quantity related to the authentication target based on the image acquired by the image acquisition means;
Storage means for storing feature quantities in association with identification information for a plurality of objects that can be authentication targets, and storing feature quantity determination information that defines how to determine feature quantities for one or more targets; ,
When feature amount determination information that defines a determination method different from a standard determination method is stored in the storage unit in association with the identification information received by the wireless communication unit, the feature amount determination Using the method of determination specified by the information, the feature quantity stored in the storage means in association with the identification information is compared with the feature quantity acquired by the feature quantity acquisition means to authenticate the authentication target. In other cases, using the standard determination method, the feature quantity stored in the storage unit in association with the identification information is collated with the feature quantity acquired by the feature quantity acquisition unit. Authentication means for authenticating the authentication target,
An image processing system comprising:

Images