JP2007094637A - Face image authentication device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce the possibility of the denial of a registerer even when performing personal authentication by a face image according to the combination of various face direction conditions and illumination conditions. <P>SOLUTION: Face featured values are calculated from a photographed face image (S211), and the similarity of the face featured values and already registered face featured values is calculated (S212), and when the similarity is larger than a similarity maximum value Pmax (S213), the similarity is set as the similarity maximum value Pmax (S214). When the similarity maximum value Pmax is not less than a first threshold Th1 (S215), the arithmetic operation (S216) of the face direction conditions of the face image and the arithmetic operation(S217) of illumination conditions are executed, and when the face featured values corresponding to the combination of the face direction conditions and the illumination conditions are not registered yet (S218), the face featured values are additionally registered by making them correspond to the combination of the face direction conditions and the illumination conditions (S219). When the similarity maximum value Pmax is less than a second threshold Th2 (S220), the photographed person is not authenticated as a registered person (S222), and when the similarity maximum value Pmax is not less than the second threshold Th2 (S220), the photographed person is authenticated as the registered person (S221). <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

The present invention relates to a face image authentication apparatus, and more particularly to a face image authentication apparatus suitable for personal authentication of a driver or the like riding in a vehicle.

Many proposals have been made to use personal authentication using a driver's face image as a vehicle security system or convenience system.

As a conventional technique for personal authentication using a face image, there is a face authentication system for ensuring security such as entrance / exit management in an office (see, for example, Patent Document 1, Patent Document 2, etc.). Application is also proposed.

JP-A-2005-115481 JP 2004-157602 A

Many of the above-described conventional authentication techniques capture facial image shading as a feature quantity that depends on an individual and compare it with a pre-registered face feature quantity to perform authentication. Etc. vary greatly depending on the combination of the face orientation condition and the lighting condition, so that if the face feature and the lighting condition combination are different from the combination of the face orientation condition and the lighting condition registered in advance, The identity of the feature amount is reduced, and a problem that authentication is rejected even by the user is generated. For example, when the pre-registered face feature amount is based on a face image taken with the face facing up, the face feature amount of the person's face image taken with the face facing down is different, The person rejection rate increases. In addition, if the face feature value registered in advance is based on a face image obtained by photographing a face under daylight, it will be different from the face feature value of the person's face image taken at dusk. The rate is high. In particular, when used in a vehicle driver's personal authentication system, the right-hand drive vehicle has to have a large amount of external light that hits the face from the right window side. There is a risk that it may differ from the facial feature amount of the face image, and there is a problem that accurate personal authentication cannot be performed.

On the other hand, if the identification threshold is set low to prevent the person from being refused authentication due to such a combination of face orientation and lighting conditions, the probability of accepting another person increases and the original personal authentication performance decreases. A malfunction that occurs.

By registering facial feature quantities under various face orientation conditions and lighting conditions at the time of initial registration, the above problems can be avoided, but various face orientation conditions and lighting conditions can be avoided. In addition, it is difficult to prepare the lighting feature in advance, and the burden on the registrant is large, so that there is a problem that it is difficult to actually perform the registration.

SUMMARY OF THE INVENTION An object of the present invention is to provide a face image authentication apparatus that automatically registers face feature amounts under various combinations of face orientation conditions and illumination conditions.

The present invention has been made in view of the above problems, and continuously captures a face image in the normal behavior of the subject, and the combination of the face orientation condition and the illumination condition in the face image is stored in the storage means. It is determined whether or not the combination of the facial orientation and the lighting condition of the facial feature amount of the registered registrant's facial image matches, and the facial orientation condition and the lighting condition that match the combination of the facial orientation condition and the lighting condition are determined. When the combination is not registered, the facial feature quantity of the face image of the subject is additionally registered in the storage unit, and the facial feature quantity of the facial image under various combinations of face orientation conditions and illumination conditions is registered. By registering facial feature quantities under various combinations of face orientation conditions and lighting conditions, there is a possibility that registrants will be rejected even when personal authentication is performed using facial images with various combinations of face orientation conditions and lighting conditions It is an object of the present invention to provide a face image authentication device that can reduce the image quality.

Means for Solving the Problems and Effects of the Invention

In order to solve the above-described problem, the face image authentication apparatus according to claim 1 captures a face feature amount of a registrant's face image for each combination of face orientation condition and illumination condition, and photographs the face of the subject. The facial feature amount of the face image of the inventor photographed by the face photographing means based on the operation of the registrant corresponding to the combination of the face orientation condition and the lighting condition at the time of photographing the face image. Face image initial registration means for initial registration in the storage means, and facial feature amounts of the face image of the subject photographed by the face photographing means triggered by a predetermined operation of the subject are registered in the storage means. When the face feature amount of the registrant's face image is similar to the face feature amount of the registrant's face image, the face feature amount of the registrant's face image is used as the face feature amount of the registrant's face image. Face image additionally registered in the storage means corresponding to the combination If the facial feature amount of the face image of the subject photographed by the additional registration means and the face photographing means is similar to any one of the facial feature amounts of the registrant face image registered in the storage means, The photographer includes authentication means for authenticating that the photographer is a registrant. According to the face image authentication device of claim 1, it is determined whether or not the face image of the registrant photographed by the face photographing camera is registered with a combination of the face orientation condition and the illumination condition registered in advance. If it is not a combination of registered face orientation conditions and illumination conditions, the registrant's face image is again registered with a combination of predetermined face orientation conditions and illumination conditions, so that the face of the subject In a system for performing personal authentication using a face image captured by a face camera, the combination of the face orientation condition and the illumination condition in the face image of the photographed person is the face orientation of the facial feature amount of the registrant face image. If the combination of the face orientation condition and the illumination condition matches the combination of the face orientation condition and the illumination condition, and the combination of the face orientation condition and the illumination condition is not registered. The face feature amount of the face image is additionally registered in the storage means, and registers the face feature amount of the face image of a combination of different face orientation conditions and lighting conditions. By registering facial feature quantities under various combinations of face orientation conditions and lighting conditions, there is a possibility that registrants will be rejected even when personal authentication is performed using facial images with various combinations of face orientation conditions and lighting conditions Can be reduced.

The face image authentication device according to claim 2 is the face image authentication device according to claim 1, wherein the face image initial registration unit includes a face feature amount calculation unit that calculates a face feature amount from a registrant's face image; A face orientation condition calculating means for calculating a face orientation condition from the person's face image, an illumination condition calculating means for calculating an illumination condition from the registrant's face image, and a combination of the face feature value of the face orientation condition and the illumination condition. And a face feature amount / face orientation condition / illumination condition storage means registered in the storage means. According to the face image authentication device of claim 2, the facial feature amount is registered in the storage unit in correspondence with the combination of the face orientation condition and the illumination condition, so that the facial feature amount is registered under any condition. It can be made clear.

The face image authentication apparatus according to claim 3, wherein the face image additional registration unit in the face image authentication apparatus according to claim 1 is a face feature amount calculation unit that calculates a face feature amount from the face image of the subject. , A face feature amount similarity comparing unit that calculates a similarity between the face feature amount calculated by the face feature amount calculating unit and the face feature amount registered in the storage unit; and the face feature amount similarity comparing unit The similarity maximum value setting means for setting the similarity as the similarity maximum value when the similarity calculated by the above is greater than the similarity maximum value, and the similarity maximum value set by the similarity maximum value setting means A face orientation condition calculating means for calculating a face orientation condition from the face image of the subject when it is equal to or greater than a first threshold; an illumination condition calculating means for computing an illumination condition from the face image of the subject; and the face orientation condition And a combination of the above lighting conditions If the facial feature value is not registered, the facial feature value calculated by the facial feature value calculation unit is registered in the storage unit in association with the combination of the face orientation condition and the illumination condition. It includes a face orientation condition / illumination condition storage means. According to the face image authentication device of claim 3, the similarity between the calculated face feature value and the face feature value registered in the storage unit is calculated, and the calculated similarity value is greater than the maximum similarity value. In this case, when the similarity is set to the maximum similarity, and the maximum similarity is equal to or greater than the first threshold, it is determined that the subject to be photographed is the original registrant and additional registration of the facial feature amount is performed. To calculate the face orientation condition of the face image and the illumination condition of the face image, and to register the face feature amount in the storage unit corresponding to the calculated combination of the face orientation condition and the illumination condition By doing so, it is possible to prevent registration of face feature values having low similarity.

The face image authentication device according to claim 4 is the face image authentication device according to any one of claims 1 to 3, wherein the authentication means authenticates the maximum similarity and the registrant's lower limit value. Is compared with the second threshold, and if the maximum similarity is equal to or greater than the second threshold, the subject is authenticated as a registrant. According to the face image authentication device of claim 4, when the maximum similarity is equal to or greater than the second threshold, the person to be imaged is authenticated as a registrant, so that the similarity with the highest certainty is used. Since the person to be photographed is authenticated as a registrant, the possibility of refusing the registrant can be reduced.

A face image authentication apparatus according to a fifth aspect is the face image authentication apparatus according to any one of the first to fourth aspects, further comprising illumination means for illuminating the face of the subject. According to the face image authentication apparatus of claim 5, since the face of the subject can be illuminated by the illumination means, a clearer face image can be obtained and the accuracy of personal authentication by the face image is further improved. Can be made.

Even when personal authentication is performed using facial images with various face orientation conditions and lighting conditions, the subject's face is photographed in a timely manner to reduce the possibility of refusing the registrant. This is achieved by additionally registering the facial feature quantity of the face image in the storage means and registering the facial feature quantity of the facial image under various combinations of face orientation conditions and illumination conditions.

Hereinafter, Example 1 of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram illustrating a face image authentication apparatus according to Embodiment 1 of the present invention. The face image authentication device according to the first embodiment includes an image processing device 10 disposed in an instrument panel corresponding to a driver's seat of a vehicle, and an LED (Light Emitting Diode) activated from the image processing device 10. A current control device 20; an LED current control device 20 connected to the LED current control device 20; and an LED 30 incorporated in an instrument panel corresponding to the driver's seat so as to illuminate a driver's face seated in the driver's seat of the vehicle; It is composed of a face photographing camera 40 such as a CCD (Charge Coupled Device) camera incorporated in an instrument panel corresponding to the driver's seat so that the driver's face seated on the seat can be photographed.

Referring to FIG. 2, the image processing apparatus 10 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12 that stores an image processing program 100, and a RAM that temporarily stores operation data of the CPU 11. (Random Access Memory) 13 and a nonvolatile memory 14 composed of an EEPROM (Electrically Erasable & Programmable Read Only Memory) or the like for registering a facial feature amount of a registrant's face image and a combination of face orientation conditions and illumination conditions. Has been.

Referring to FIG. 3, the image processing program 100 includes a face image registration start signal input means 101, a face image registration end signal input means 102, a shooting trigger input means 103, a registration condition reading means 104, and other devices. Communication means 105, image input means 106, face feature quantity calculation means 107, face feature quantity similarity comparison means 108, first threshold value storage means 109, second threshold value storage means 110, face orientation condition The calculation unit 111, the illumination condition calculation unit 112, the face feature amount / face orientation condition / illumination condition storage unit 113, and the registered condition similarity comparison unit 114 are configured.

FIG. 4 is a flowchart showing face image initial registration processing (face image initial registration means).

FIG. 5 is a flowchart showing face image addition registration processing (face image addition registration means). In the first embodiment, as a registrant determination threshold for personal authentication processing using a face image, a first threshold Th1 that is a lower limit value of similarity that permits additional registration of a facial image, and a registrant in personal authentication using a facial image, It has a second threshold value Th2 that is a lower limit value of the similarity to be determined, and the first threshold value Th1> the second threshold value Th2 (assuming that a larger threshold value is more severe). For example, the first threshold value Th1 is set to be a registrant rejection rate of 0.5%, and the second threshold value Th2 is set to be a registrant rejection rate of 2%.

FIG. 6 is a conceptual diagram illustrating a combination of face orientation conditions and illumination conditions for registering face feature amounts in the nonvolatile memory 14. Under the four lighting conditions of uniform lighting, bright lighting on the right face, bright lighting on the left face, and bright lighting on the lower half of the face, the upper and lower faces are each -30 ° to + 30 ° in 10 ° increments and left and right faces. However, the facial feature amount is registered every 10 ° from −30 ° to + 30 °. In this embodiment, in order to reduce the amount of information to be registered, (−30 °, −30 °, − (30 °), (−30 °, + 30 °), (+ 30 °, −30 °), and (+ 30 °, + 30 °) four conditions near the face direction are excluded.

Next, the operation of the face image authentication apparatus according to the first embodiment configured as described above will be described with reference to the flowcharts of FIGS. 4 and 5.

(1) Face image initial registration process (see Fig. 4)

In carrying out the present invention, a facial feature amount of a registrant's face image in a combination of at least one face orientation condition and illumination condition and a combination of face orientation condition and illumination condition at the time of face image shooting are initially registered. Therefore, the image processing apparatus 10 first executes a face image initial registration process shown in FIG.

When a person seated in the driver's seat of the vehicle (referred to as a registrant in the face image initial registration process) operates a face image registration button or the like to register his / her face image, the image processing apparatus 10 receives a face image registration start signal. A face image registration activation signal is input by the input means 101 (S101 in FIG. 4), and the image processing program 100 is activated (S102 in FIG. 4). The face image registration activation signal input unit 101 includes an authentication unit such as a personal identification number, and starts the image processing program 100 based on the authentication result.

Next, the image processing apparatus 10 activates the LED current control apparatus 20 (S103 in FIG. 4). Thereby, light emission of LED30 is attained.

Subsequently, the image processing apparatus 10 activates the face photographing camera 40 (S104 in FIG. 4). This makes it possible to photograph the registrant's face.

Next, when the registrant operates the shooting button or the like, the image processing apparatus 10 inputs a shooting trigger signal by the shooting trigger signal input means 103 (S105 in FIG. 4), and simultaneously causes the LED 30 to emit light, and at the same time, a face shooting camera. The face image of the registrant is photographed by 40 (S106 in FIG. 4). The photographed face image of the registrant is input from the face photographing camera 40 to the image processing apparatus 10 by the image input means 106 and temporarily stored in the RAM 13.

Subsequently, the image processing apparatus 10 calculates the face feature amount from the face image temporarily stored in the RAM 13 by the face feature amount calculation unit 107 (S107 in FIG. 4). Specifically, the face feature amount computing unit 107 first detects the registrant's face area from the face image temporarily stored on the RAM 13. For example, by obtaining a correlation value while moving a template prepared in advance in a face image, a place having the highest correlation value is set as a face region. In addition, a face region extraction method using an eigenspace method or a subspace method may be used. Next, the face feature amount calculation means 107 detects the position of the face parts such as eyes, nose, and mouth from the detected face area. As this detection method, for example, literature (Kazuhiro Fukui, Osamu Yamaguchi: “Face feature point extraction by combination of shape extraction and pattern matching”, IEICE Transactions (D), vol. J80-D-II, No. 8, pp2170-2177 (1997)). Subsequently, the face feature quantity computing means 107 cuts out the face area into a certain size and shape based on the detected position of the face part, and uses the shading information as the face feature quantity. For example, the gray value of an area of m pixels × n pixels can be used as information as it is, and m × n-dimensional information can be used as a feature vector.

Next, the image processing apparatus 10 calculates the face orientation condition from the face image temporarily stored in the RAM 13 by the face orientation condition calculating unit 111 (S108 in FIG. 4). Specifically, the face orientation condition calculation unit 111 checks the face orientation based on the face detection based on the face template and the relative position of the detection coordinates of the face parts such as eyes and nose.

Subsequently, the image processing apparatus 10 calculates the illumination condition from the face image temporarily stored in the RAM 13 by the illumination condition calculation unit 112 (S109 in FIG. 4). For example, the illumination condition calculation unit 112 divides the face image into four image areas of an upper left quadrant, an upper right quadrant, a lower left quadrant, and a lower right quadrant, and compares the average brightness values of the image areas. It is determined which of the four illumination conditions is uniform illumination, illumination with a bright right face, illumination with a bright left face, and illumination with a bright lower half of the face.

Next, the image processing apparatus 10 uses the face feature amount / face orientation condition / illumination condition storage unit 113 to convert the face feature amount calculated in step S107 to the face orientation condition calculated in step S108 and the illumination calculated in step S109. The information is registered in the nonvolatile memory 14 corresponding to the combination of conditions (S110 in FIG. 4). At this time, the registered condition similarity determination unit 114 causes the face orientation condition and the illumination condition closest to the calculated face orientation condition and the calculated illumination condition combination from the combinations of the face orientation condition and the illumination condition shown in FIG. Is selected.

Then, the image processing apparatus 10 inquires the registrant whether or not to end the initial registration of the face image (S111 in FIG. 4), and sends a face image registration end signal based on the operation of the registrant's face image registration end button or the like. Unless otherwise input by the face image registration end signal input means 102 (S111: No in FIG. 4), control is returned to step S105, and steps S105 to S111 are repeated.

When the registrant operates the face image registration end button or the like after performing registration corresponding to the combination of the face direction and the lighting condition of the face feature quantity one or more times, the image processing apparatus 10 displays the face image registration end signal input means. The face image registration end signal is input at 102 (S111: Yes in FIG. 4), and the face image initial registration process is ended (S112 in FIG. 4).

(2) Face image additional registration processing (see FIG. 5)

The image processing apparatus 10 executes a face image additional registration process in response to a person's intention to board the vehicle.

When a person (referred to as a photographed person in the face image additional registration process) gets into the driver's seat of the vehicle, the seating of the photographed person in the driver's seat is detected by a sensor (such as a load sensor) built in the driver's seat ( 5 (S201 in FIG. 5), the image processing apparatus 10 activates the image processing program 100 (S202 in FIG. 5). Instead of detecting the seating of the photographed person in the driver's seat, a face image registration activation signal may be input when the vehicle door is unlocked. As described above, it is a feature of the present invention that the image processing program 100 is automatically started according to the user's intention to board the driver's seat without an instruction from the user.

Next, the image processing apparatus 10 reads the combination of the face orientation condition and the illumination condition for which the face feature amount is not registered from the nonvolatile memory 14 by the registered condition reading unit 104 (S203 in FIG. 5).

Subsequently, the image processing apparatus 10 calculates the number of combinations of face orientation conditions and illumination conditions for which facial feature quantities have already been registered (condition registration count) (S204 in FIG. 5), and sets the face orientation conditions and illumination conditions. It is determined whether one or more combinations are registered (S205 in FIG. 5).

When there is no combination of the face orientation condition and the illumination condition in which the facial feature amount is registered (S205: No in FIG. 5), the image processing apparatus 10 uses the communication means 105 to other devices to start the initial facial image. After notifying the registration process (see FIG. 4) to other devices (for example, a meter device, a car navigation device, a mobile phone, etc.) and notifying the photographed person, vehicle owner, etc. (see FIG. 5) S226), the face image additional registration process is terminated (S227 in FIG. 5).

If there is one or more combinations of face orientation conditions and illumination conditions for which facial feature amounts are registered (S205 in FIG. 5: Yes), the image processing apparatus 10 registers the facial feature amounts from the nonvolatile memory 14. In addition, the combination of the face orientation condition and the illumination condition is read (S206 in FIG. 5).

Next, the image processing apparatus 10 initially sets the similarity maximum value Pmax in the personal authentication operation using the current face image to 0 (S207 in FIG. 5).

Subsequently, the image processing apparatus 10 activates the LED current control apparatus 20 (S208 in FIG. 5). Thereby, light emission of LED30 is attained.

Next, the image processing apparatus 10 activates the face image capturing camera 40 (S209 in FIG. 5). As a result, the face of the subject can be photographed.

Subsequently, the image processing apparatus 10 automatically inputs a shooting trigger signal by the shooting trigger input means 103 to cause the LED 30 to emit light, and simultaneously operates the shutter of the face image shooting camera 40 to capture the face image of the subject. A picture is taken (S210 in FIG. 5). The photographed face image of the photographed person is input from the face photographing camera 40 to the image processing apparatus 10 by the image input means 106 and temporarily stored in the RAM 13.

Next, the image processing apparatus 10 calculates the face feature amount from the face image of the person to be photographed temporarily stored in the RAM 13 by the face feature amount calculation unit 107 (S211 in FIG. 5). The calculation of the facial feature amount has already been described in detail in step S107.

Subsequently, the image processing apparatus 10 calculates the degree of similarity between the face feature amount calculated in step S211 and the face feature amount initially registered in the nonvolatile memory 14 by the face feature amount similarity comparing unit 108. (S212 of FIG. 5). Specifically, the face feature amount similarity comparison unit 108 calculates the similarity of the face feature amount using a method such as a subspace method or a composite similarity method. For example, literature (Kenichi Maeda, Sadaichi Watanabe: “Pattern matching method introducing local structure”, IEICE Transactions (D), vol. J68-D, No. 3, pp 345-352 (1985)). The mutual subspace method described in can be used. In this method, both the calculated face feature value and the initially registered face feature value are expressed as a partial space, and an “angle” formed by the two partial spaces is defined as a similarity. Of course, the degree of similarity may be defined by another scale such as a correlation value or Euclidean distance.

Next, the image processing apparatus 10 determines whether or not the calculated similarity is larger than the maximum similarity Pmax (S213 in FIG. 5), and if it is larger (S213 in FIG. 5: Yes), the calculated similarity is determined. The degree is set to the similarity maximum value Pmax (S214 in FIG. 5). If the calculated similarity is equal to or less than the maximum similarity Pmax (S213: No in FIG. 5), step S214 is skipped.

Subsequently, the image processing apparatus 10 compares the maximum similarity value Pmax and the first threshold value Th1 that is stored in the first threshold value storage unit 109 and that is a lower limit value of the similarity level that permits additional registration of face images. (S215 in FIG. 5).

When the similarity maximum value Pmax is equal to or greater than the first threshold value Th1 (S215: Yes in FIG. 5), the image processing apparatus 10 moves control to step S216, and the face orientation condition calculation unit 111 determines the face image of the subject. Calculation of face orientation conditions (S216 in FIG. 5: the same as S108 in FIG. 4), calculation of illumination conditions of the face image of the subject by the illumination condition calculation means 112 (S217 in FIG. 5: similar to S109 in FIG. 4) It is executed, and it is determined whether or not the face feature amount is unregistered corresponding to the calculated combination of the face orientation condition and the illumination condition (S218 in FIG. 5). When the face feature amount is not registered corresponding to the combination of the face orientation condition and the illumination condition (S218 in FIG. 5: Yes), the image processing apparatus 10 stores the face feature amount / face orientation condition / illumination condition storage unit 113. Thus, the facial feature quantity of the face image of the subject is registered in the nonvolatile memory 14 in association with the combination of the face orientation condition and the illumination condition (S219 in FIG. 5), and the control is shifted to step S220. At this time, the registered condition similarity determination unit 114 causes the face orientation condition and the illumination condition closest to the calculated face orientation condition and the calculated illumination condition combination from the combinations of the face orientation condition and the illumination condition shown in FIG. Is selected. When the facial feature amount has already been registered corresponding to the combination of the face orientation condition and the illumination condition (S218: No in FIG. 5), the image processing apparatus 10 sets the facial feature amount of the registrant's face image to nonvolatile. Without registering in the memory 14, control is passed to step S220.

On the other hand, when the similarity maximum value Pmax is less than the first threshold value Th1 (S215: No in FIG. 5), the image processing apparatus 10 shifts the control to step S220 without executing steps S216 to S219.

In step S <b> 220, the image processing apparatus 10 stores the similarity maximum value Pmax and the second threshold value that is stored in the second threshold value storage unit 110 and is the lower limit value of the similarity level that authenticates the registrant in personal authentication using a face image. Compare with Th2.

When the maximum similarity Pmax is less than the second threshold Th2 (S220: No in FIG. 5), the image processing apparatus 10 determines that the face image of the subject is a face image other than the registrant, that is, Without authenticating that the subject is a registered person (S222 in FIG. 5), the authentication result is transmitted to another device (S223 in FIG. 5).

On the other hand, when the maximum similarity Pmax is equal to or greater than the second threshold Th2 (S220 in FIG. 5: Yes), the image processing apparatus 10 determines that the face image of the subject is the face image of the registrant, That is, the person to be photographed is authenticated as a registrant (S221 in FIG. 5), and the authentication result is transmitted to another device (S223 in FIG. 5).

Next, the image processing apparatus 10 determines whether or not the maximum similarity value Pmax is equal to or greater than the second threshold Th2 and the same subject is continuously seated in the driver's seat (S224 in FIG. 5). Whether or not the same subject continues to be seated in the driver's seat is determined by, for example, detecting that the subject is not away by a load sensor built in the seat.

When it is determined that the maximum similarity value Pmax is equal to or greater than the second threshold Th2 and the same subject is continuously seated in the driver's seat (S224: Yes in FIG. 5), the image processing apparatus 10 After the calculation of the shooting interval (S225 in FIG. 5), control is returned to step S210, and steps S210 to S224 are executed again. Thus, the face images can be automatically and sequentially registered in correspondence with combinations of face orientation conditions and illumination conditions for which the face images are not registered.

When the similarity maximum value Pmax is equal to or greater than the second threshold Th2 and it is not determined that the same subject is continuously seated in the driver's seat (S224: No in FIG. 5), the image processing apparatus 10 Finishes the face image addition registration process (S227 in FIG. 5).

By repeating the series of operations as described above, registrants automatically add facial feature quantities to α in various combinations of face orientation conditions and lighting conditions without repeating troublesome facial image registration operations. You can register. For this reason, when performing personal authentication using a face image, authentication is performed based on a plurality of automatically registered facial feature amounts. The possibility of rejecting the registrant can be reduced even if the combination of lighting conditions is slightly different.

In the first embodiment, the description has been made on the premise of a system for performing personal authentication using a face image obtained by photographing the face of the driver of the vehicle with the face photographing camera 40. However, the application target of the present invention is a vehicle. Needless to say, it is not limited to personal authentication. For example, the present invention can be applied in various fields such as an entrance management system for libraries and museums, an entrance / exit management system for companies, laboratories, etc., and a personal authentication system for camera-equipped mobile phones.

As mentioned above, although the Example of this invention was described, this is only an illustration, this invention is not limited to this, Based on the knowledge of those skilled in the art, unless it deviates from the meaning of a claim Various changes are possible.

1 is a block diagram showing a configuration of a face image authentication device according to Embodiment 1 of the present invention. The circuit block diagram which shows the structure of the image processing apparatus in FIG. The block diagram which shows the structure of the image processing program in FIG. 6 is a flowchart illustrating face image initial registration processing in the face image authentication apparatus according to the first embodiment. 5 is a flowchart illustrating face image addition registration processing in the face image authentication apparatus according to the first embodiment. The figure which shows the example of the combination of the face direction conditions and illumination conditions which register the face image in the face image addition registration process of FIG.

Explanation of symbols

10 Image processing device 11 CPU
12 ROM
13 RAM
14 Nonvolatile memory 20 LED current control device 30 LED
40 face photographing camera 100 image processing program 101 face image registration start signal input means 102 face image registration end signal input means 103 photographing trigger input means 104 registration condition reading means 105 communication means 106 to other equipment 106 image input means 107 facial features Quantity calculation means 108 Face feature quantity similarity comparison means 109 First threshold storage means 110 Second threshold storage means 111 Face orientation condition calculation means 112 Illumination condition calculation means 113 Facial feature quantity / face orientation condition / illumination condition storage means 114 Registration conditions Similarity comparison means

Claims (5)

Storage means for registering the facial feature amount of the registrant's face image for each combination of face orientation condition and illumination condition;
A face photographing means for photographing the face of the subject,
The facial feature quantity of the inventor's face image photographed by the face photographing means based on the operation of the registrant is initially registered in the storage means in correspondence with the combination of the face orientation condition and the illumination condition at the time of photographing the face image. Face image initial registration means;
When the facial feature quantity of the face image of the photographed person photographed by the face photographing means triggered by the predetermined action of the photographed person is similar to the facial feature quantity of the registrant face image registered in the storage means Further, the face feature amount of the face image of the photographed person is additionally registered in the storage unit as the face feature amount of the registrant face image in correspondence with the combination of the face orientation condition and the illumination condition at the time of face image shooting. Image additional registration means;
When the facial feature amount of the face image of the subject photographed by the face photographing means is similar to any one of the facial feature amounts of the registrant face image registered in the storage means, the subject is a registered person An authentication means for authenticating that
A face image authentication apparatus comprising: The face image initial registration unit calculates a facial feature amount from a registrant's face image, a facial direction condition calculation unit that calculates a facial direction condition from the registrant's face image, and a registrant's face Illumination condition calculation means for calculating an illumination condition from an image, and facial feature quantity / face orientation condition / illumination condition storage means for registering the face feature quantity in the storage means in association with the combination of the face orientation condition and the illumination condition The face image authentication apparatus according to claim 1, comprising: The face image addition registration means is registered in the face feature amount calculating means for calculating the face feature amount from the face image of the subject, the face feature amount calculated by the face feature amount calculating means, and the storage means. A face feature quantity similarity comparison unit that calculates a similarity with a face feature quantity, and if the similarity calculated by the face feature quantity similarity comparison unit is greater than the maximum similarity value, the similarity is set to the maximum similarity value And a face orientation for calculating a face orientation condition from the face image of the subject when the similarity maximum value set by the similarity maximum value setting means is equal to or greater than a first threshold. A condition calculation unit, an illumination condition calculation unit that calculates an illumination condition from a face image of the subject, and the face feature when the face feature amount is not registered corresponding to a combination of the face orientation condition and the illumination condition. Facial features calculated by quantity calculation means A face image authentication device according to claim 1 or 2, wherein including said face orientation condition and the facial feature quantity and face orientation conditions and lighting condition storage means for registering in the storage means in association with the combination of the illumination conditions. The authentication means compares the maximum similarity value with a second threshold value that is a lower limit value of the similarity level that authenticates the registrant. If the maximum similarity value is equal to or greater than the second threshold value, The face image authentication device according to claim 1, wherein the face image authentication device authenticates a registrant. The face image authentication apparatus according to claim 1, further comprising an illumination unit that illuminates the face of the subject.

Images