JP2004318501A - Figure authentication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To limit the use of a desired system according to human authentication results in an environment where the desired system is shared by a number of people. <P>SOLUTION: A person image P of the person who is to operate a personal computer or the like is picked up by an image pickup means and taken in to an image data input part 2. A face area extracting part 3 extracts only the face area from the person image and a feature quantity detecting part 4 samples a plurality of predetermined points within the face area to detect feature quantities. The feature quantities are computed by a feature quantity computing imaging part 5 into feature quantity data showing the positions and sizes of the eyebrows, eyes, nose, mouth and the like. Feature quantity reference value data obtained by averaging the feature quantities of a number of people are stored in a feature quantity recording part 6 in advance. An authenticating part 7 compares the feature quantity data to the feature quantity reference value data to authenticate the person, with the authentication result reported to an authentication result part 8. An instruction to limit the use of the personal computer or the like is given according to the person authenticated. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a person authentication device that performs authentication of a person who intends to perform a desired operation of a device, and more specifically, performs feature authentication by extracting features from an image of a person's face. The present invention relates to a person authentication device that appropriately instructs an operation corresponding to a desired device such as a computer device according to an authentication result.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, when a user of a personal computer (hereinafter, referred to as a personal computer) is specified, personal authentication by password registration or the like has been performed. To do so, it is necessary to set password registration for each user who uses a personal computer. Patent Document 1 below discloses a technology of a computer system capable of performing personal authentication based on an image of a user's face. According to this technique, when the user operates the personal computer, the user's face is photographed by a camera provided at or near the personal computer, and the face image acquired by the photographing is compared with a previously registered face image. To perform person authentication. When a specific person is authenticated, a predetermined process is executed, thereby realizing person authentication of the computer system. Accordingly, it is possible to automatically and continuously execute arbitrary processing such as activation of an application set for each user in a batch process in accordance with the result of person authentication.
[0003]
[Patent Document 1]
JP 2001-14052 A (paragraph numbers 0020 to 0058, FIGS. 1 to 13)
[0004]
[Problems to be solved by the invention]
However, in the conventional person authentication device used for the computer system including the technique of the above-mentioned Patent Document 1, a face image is newly registered as data for a person whose face data is not registered in advance. Without it, no person authentication can be performed. Furthermore, if you try to run an application according to age, gender, etc. on a personal computer used in an environment that is supposed to be used by an unspecified number of people, the execution of that application must be attributed to each individual. It does not support performing authentication that restricts the operation according to. In addition, after authenticating a predetermined person at the time of starting up a personal computer, there is no use restriction when a different person is replaced halfway. Not only personal computers, but also when performing personal authentication when trying to operate a desired device, the same problem as described above occurs in a conventional personal authentication device.
[0005]
The present invention has been made in view of such circumstances, and in an environment where a specific person shares a desired device or an unspecified number of people use the desired device, the authentication result of the person or the user is used. It is an object of the present invention to provide a person authentication device capable of coping with a use restriction of a desired device according to a result of the determination. More specifically, in a computer system environment where a specific person shares a personal computer or an unspecified number of persons use a personal computer, the use of personal computers is restricted according to the result of personal authentication or the result of user identification. It is an object of the present invention to provide a person authentication device that can take the following.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, the person authentication device according to the present invention captures an image of a person who is to perform an operation before starting use of a desired device such as a personal computer by an imaging unit. Then, a face image region is extracted from the captured image of the person, and sampling data such as the position, size, and color of each predetermined portion such as eyebrows, eyes, mouth, and nose in the face image region is detected as a feature amount. Then, this feature amount is calculated and held as feature amount data of a person to be authenticated. On the other hand, feature amount reference value data obtained by averaging feature amount data of many persons is held in advance, and the feature amount data of the person to be authenticated and the feature amount reference value data are compared and analyzed. For example, the attribute of a person who intends to execute a personal computer operation is specified based on the attribute. The attribute of a person may be a characteristic division of gender such as a man and a woman, or may be a characteristic division of age such as a child and an adult or an adult other than the elderly. Then, an instruction is given to determine a range in which a desired device such as a personal computer can be handled according to the attribute of the specified person. In this way, it is possible to restrict the use of a desired device such as a personal computer according to the result of authentication by the person authentication device or make it easier to use, so that each person can use a personal computer etc. in the optimal environment The range can be changed. Here, the “corresponding range” is a concept including various usage patterns and usage modes (modes, types of reproduction / recording, number of times, etc.), including whether or not the device can be used. is there.
[0007]
That is, according to the present invention, the image of the face of the person is extracted from the image of the person obtained by imaging the person who intends to perform the operation of the desired device, and the authentication of the person and the A person authentication device for performing an instruction of a responsive range,
Face area extraction means for capturing the image of the person and extracting an image area of the face,
A feature amount detecting unit that detects, as a feature amount, sampling data for each predetermined portion of the face image region extracted by the face region extracting unit;
A feature amount calculation image processing unit that calculates the feature amount detected by the feature amount detection unit to create feature amount data;
Feature amount recording means for recording feature amount reference value data previously obtained from feature amount data of a plurality of persons;
Comparing and analyzing the feature amount data obtained from the feature amount calculation image processing unit and the feature amount reference value data obtained from the feature amount recording unit,
Identifying the attribute of the person based on the analysis result,
Setting a possible range of the desired device according to the specified attribute,
An authentication unit that instructs the device to indicate the applicable range,
A personal authentication device is provided.
[0008]
In addition, the face region extracting means extracts the image of the face by capturing the image of the person every predetermined time or every time a predetermined change occurs in the attribute of the person,
It is a preferable aspect of the person authentication apparatus according to the present invention that the authentication unit compares and analyzes the feature amount data and the feature amount reference value data each time the face area detection unit extracts a face image area. That is, according to the person authentication device of the present invention, when a desired device such as a personal computer is operated, the operating person is imaged at predetermined time intervals, for example, every 10 minutes or every 30 minutes. Quantitative data is extracted. Alternatively, when a predetermined change occurs in the attribute of the person who operates, such as when a person operating a desired device such as a personal computer changes, the person is imaged to extract feature amount data. Then, each time the feature amount data is extracted, the feature amount reference value data registered / stored in advance is compared with the feature amount data to determine whether or not they match, and authentication is performed each time. As a result, it is possible to prevent an unauthorized operation from being performed by a person having a different attribute than when a desired device such as a personal computer is started.
[0009]
BEST MODE FOR CARRYING OUT THE INVENTION
The present invention provides a person authentication device that determines a person based on the features of an image of a person's face and gives an instruction to execute a desired device operation for each determined person. In the following embodiments, for the sake of simplicity, a personal computer is used as a desired device, and a person authentication device capable of coping with an appropriate computer system environment for each person determined by the person authentication device. Will be described. That is, the image of the face of a person taken by the image pickup means and analyzed by the image input means is analyzed to determine the person from the attributes of the person, such as an old person, a young person, a man, and a woman. An appropriate command is transmitted from the person authentication device to the personal computer so that the operation environment of the personal computer is prepared. As a result, the operability and convenience in handling the personal computer of the person who has been authenticated by the person authentication device can be further improved.
[0010]
The person authentication device of the present invention captures a person's face by an image pickup means at the start of use of the personal computer, and registers and accumulates the data as image data (face data). There is provided a judging means for collating face data of a person to be authenticated with registered and accumulated image data to judge whether or not they match. Furthermore, in the case where face data is not registered and stored, or in a computer system that does not require registration of face data, the image data (face data) captured by the imaging means is independently analyzed to identify the old, young, male, Rough person authentication is performed, such as for women, so that personal computers can respond accordingly.
[0011]
Here, the analysis of the face data is performed as follows. First, an average value of feature amounts is obtained from face data of many adult men. The feature amount is sampling data such as the position, size, and color of each predetermined portion such as eyebrows, eyes, nose, and mouth in the face image. For example, an average value of feature amounts is obtained using face data in registration data of 1,000 adult men. The average value of the feature values of the adult male face data is used as the feature value reference value of the adult male. Similarly, feature amount reference values are obtained from face data for many adult women. Then, by comparing the feature amount of the newly acquired face data and the feature amount reference value to perform the person authentication, the person identification is performed by identifying the person while determining attributes such as old / young / man / woman. Do.
[0012]
The basic criteria for judging elderly men, elderly women, adult men, adult women, adolescent men (boys), adolescent women (girls), boys and girls are listed below. "Adult" excludes "elderly". Adult women are most likely to be wearing makeup when they go out, especially when applying lipstick to their lips. Therefore, when sampling the color of the lips from the face data to determine the distribution of the color components, if there is a bias in a specific color component such as a large red component and it differs from the original lip color, to decide. Conversely, if the result of sampling the color of the lips from the face data has no significant difference from the distribution of the color components of the original lips of a human, it is determined to be a man or a child. Furthermore, if there is a throat and a beard, it is determined that the subject is an adult man. However, even if the throat and the beard cannot be confirmed, it is not always possible to determine that the subject is not an adult man. Lip color, etc.).
[0013]
Also, if the hairstyle is remarkably short (for example, shaved head, square cut, sports cut, or skin head), the hair may be regarded as a man. Furthermore, if the earrings are worn as earrings larger than the earrings, the woman may be determined to be a woman. In addition, the forehead, the corners of the eyes, the cheeks, the wrinkles at the mouth, and the like are sampled, and if the feature amount of these sampled data is larger than the feature amount reference value, it is determined that the person is an elderly person. Also, comparing the face of an adult with the face of a child, the size of the face is naturally smaller for the face of the child, the components such as eyes and mouth are smaller, and the distance between the components such as eyes and mouth is also smaller. Since the child's face is shorter, if these components are smaller than the feature amount reference value or the distance is short, it is determined that the child is a boy such as a boy or a girl. Furthermore, since the skeleton of a child is less mature than that of an adult, it is possible to make a more accurate determination between an adult and a child by weighting the contour of the face and the degree of development of the jaw bone.
[0014]
The difference in skeleton between men and women occurs after puberty. However, since there is no difference in skeleton between boys and girls, discrimination between boys and girls based on facial contours is avoided. Furthermore, since there is little difference in the skeleton between adult women, adolescent men (boys), and adolescent women (girls), discrimination between them based on facial contours and the like is avoided. In addition, the features of a female face are more rounded than those of a male face, and the ratio of the distance between eyes to the size of the face is greater in a female face than in a male face. In other words, if the ratio of the distance between the eyes to the size of the face is small, the face looks sharp like a male face. Therefore, it is possible to determine an adult man and an adult woman based on the feature amount of the roundness of the contour of the face and the ratio of the distance between the eyes to the size of the face. There are many exceptions for the density, length, and black and white of the hair, and therefore, they are not used as criteria for judging elderly people, adults, and men and women.
[0015]
Hereinafter, some embodiments of the person authentication device according to the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram of a person authentication apparatus according to the present invention, which fetches an image of a person's face by an image data input unit and compares the image with a registered face image to authenticate a person. In FIG. 1, a person authentication device 1 includes an image data input unit 2 for capturing data of a person image, and a face area for extracting only a necessary face image region from the data of the person image captured by the image data input unit 2. An extraction unit (face area extraction unit) 3; and a feature amount detection unit (feature amount detection unit) that performs sampling for each predetermined portion of the face in order to extract the characteristics of a person from the face region extracted by the face area extraction unit 3. 4) a feature amount calculation image processing unit (feature amount calculation image processing means) 5 for calculating feature position data by calculating the position of the points sampled by the feature amount detection unit 4 and the relative distance between the points; A feature value recording unit (feature value recording means) 6 for recording feature value reference value data relating to an image of a registered person's face; and feature value data created by the feature value calculation image processing unit 5 An authentication unit (authentication means) 7 for comparing the feature amount reference value data recorded in advance in the amount recording unit 6 to authenticate a person; and acquiring the result of authentication by the authentication unit 7 and notifying the personal computer. It comprises an authentication result unit 8 and a control unit 9 for controlling the entire person authentication device 1.
[0016]
<First embodiment>
First, a person authentication device according to the first embodiment of the present invention will be described. When the personal authentication device according to the first embodiment starts up a personal computer having an image sensor such as a CCD, an image of a face of a person sitting in front of the personal computer is photographed by the image sensor such as a CCD and registered in the personal computer in advance. The system has a system configuration in which the data is compared with the data of the face image that has been set, and it is determined whether or not they match. At this time, the algorithm for extracting and authenticating the face image may be a comparative authentication technique generally used in the related art, or may be a method for detecting a detected face as described in, for example, JP-A-6-259534. An algorithm for performing authentication by comparing the feature amount of the area with the feature amount for authentication may be used.
[0017]
The operation of the person authentication device will be described based on FIG. When a person tries to operate a personal computer (not shown), an image of the person to be authenticated (person image P) is captured by an image sensor (not shown). Then, the person image P photographed by the image sensor is input to the image data input unit 2. At this time, when the input person image P includes information other than the face area such as the whole body / upper body image, the face area extracting unit 3 extracts only the image of the face area (face image). FIG. 2 is an image diagram of a face image extracted by the face area extraction unit 3 in FIG. That is, an image of a face as shown in FIG.
[0018]
Next, in order to extract the features of the person from the face image as shown in FIG. 2, the feature amount detection unit 4 performs sampling for each face portion. FIG. 3 is an image diagram for extracting a feature amount of face data from the face image shown in FIG. In the example of FIG. 3, for the eyebrow 10, the tip of the eyebrow, the tail of the eyebrow, and the middle part thereof are sampled. For the eye 11, the part of the inner eye, the corner of the eye, the middle, the upper eyelid, and the lower eyelid are sampled. As for 12, sampling is performed at several points in the area. In addition, the face contour / jaw 13 is sampled at several points along the contour. Further, for the mouth 14, the mouth corner, the upper lip, the lower lip, the opening, and the like are sampled. With respect to other face portions, a predetermined portion can be sampled if it is set in advance. The sampling data for each part obtained in this way is defined as a feature value.
[0019]
Each feature amount obtained as described above is subjected to image processing by the feature amount calculation image processing unit 5. That is, after various processes such as brightness adjustment, noise removal, and edge enhancement are performed by the feature amount calculation image processing unit 5, the coordinates of the sampling points are obtained, and then the distance between the sampling points (relative distance) and the like are obtained. Is performed to create feature amount data. Next, the authentication unit 7 compares the feature amount data created by the feature amount calculation image processing unit 5 with the feature amount reference value data of the face image registered in the feature amount recording unit 6 in advance. Is determined.
[0020]
When comparing the feature amount data with the feature amount reference value data, an arbitrary threshold value s (%) is set for the correlation between the two. If the result of the comparison by the authentication unit 7 is that the feature amount data is equal to or larger than the threshold value s (%), information of “match” is transmitted to the authentication result unit 8. Thus, the authentication result unit 8 authenticates the person who is going to be authenticated, and notifies the personal computer (not shown) that the operation is permitted.
[0021]
If the authentication unit 7 compares the feature amount data with the feature amount reference value data and finds that the feature amount data is smaller than the threshold value s (%), the control unit 9 controls the photographing conditions of the person. Is changed, and the person image P is again input to the image data input unit 2. In this way, the re-input operation of the person image P is repeated a certain number of times, and if the comparison between the feature data and the feature reference value data does not match in the authentication unit 7, the information of "mismatch" is sent to the authentication result unit 8. Tell In this way, if the feature amount data of the person that has just been captured is too far from the feature amount reference value data registered in advance, the person who is going to be authenticated at this determination stage is required to operate the personal computer. After making a final determination that the person is not a person to whom the right has been granted, the authentication result unit 8 notifies the personal computer that the operation is not permitted. By such a flow of the authentication operation, the person registered in advance is almost specified as the person who requests permission to operate the personal computer. If necessary, the feature amount data registered in the feature amount recording unit 6 may be updated when the authentication unit 7 performs the authentication comparison.
[0022]
<Second embodiment>
Next, a person authentication device according to a second embodiment of the present invention will be described. The personal authentication apparatus according to the second embodiment performs personal authentication when a single personal computer is shared by all the family members at home and gives each person an operation right. In a conventional personal computer using a person authentication unit by password registration, for example, a child can log in using a father's password. In this case, the current user is recognized as the father (adult) by the personal computer, so if the child himself connects to the Internet, he / she will see sites that are not suitable for children (sex, violence, etc.) without any restrictions. Can be done. In this case, you will be asked, "Are you 18 years old or older?", But if the child trying to operate the computer clicks "Yes", unlimited connections will be made to sites that are not suitable for children. There is virtually no regulation.
[0023]
Therefore, using the same algorithm as that in the first embodiment, the face data of the person image P captured by the image data input unit 2 of the person authentication device 1 shown in FIG. The shape, position, etc. of are extracted. Since the skeleton of a child is less mature than an adult, a difference appears in the outline of the face and the development of the jawbone. Naturally, since the size of the face is smaller in a child, components such as eyes and mouth are also small, and the distance between the components is short.
[0024]
FIG. 4 is an image diagram in which the person authentication device 1 shown in FIG. 1 extracts a feature amount of face data from an image of a face taken from an image of an adult person. FIG. 5 is an image diagram of extracting a feature amount of face data from a face image taken from a child person image in the person authentication device 1 shown in FIG. As can be seen by comparing FIGS. 4 and 5, even if the feature amount of the same face data portion is extracted from the image of the adult face in FIG. 4 and the image of the child's face in FIG. Since the relative distances and the positions of the sampling points, such as the outline of the face and the outline of the chin 13 and the mouth 14, are different, the authentication unit 7 can easily determine the difference between an adult and a child.
[0025]
A specific example of detecting the relative distance between points and the position of points in the face data of an adult and the face data of a child will be described in further detail. FIGS. 6A and 6B are conceptual diagrams when extracting the feature amount of the face data in the person authentication device 1 shown in FIG. 1, and FIG. 6A shows a part of the face data. FIG. 6B illustrates extraction of a feature amount of another part of the face data. As shown in FIG. 6A, the length of the eyebrows (between the eyebrows to the outside of the eyebrow) a and the height of the eyebrows b, the width of the eyes (between the eyes and the outside of the eyes) c and the vertical width d of the eyes, the width of the mouth e and the width of the mouth For the height f, the length g of the nose, and the like, the feature amount reference value of the adult male and the feature amount of the child are compared. Also, as shown in FIG. 6B, the interval between both eyes (left eye to right eye) h, the interval between both eyebrows (left eye to right eyebrows) i, the distance between the nose and the eye (nose to eye) j, the nose Distance between nose and mouth (nose-to-mouth angle) k, distance between nose and eyebrow (nose-to-eyebrow) m, distance between nose and chin (nose-to-chin) n, face width p, chin-to-chin (ella) As for the angle α and the angle β of the face outline, the feature amount reference value of the adult male and the feature amount of the child are compared.
[0026]
Depending on the child to be authenticated, there may be a case where a part of these features is not much different from the feature amount reference value of the adult male. Therefore, one of the following two determination methods is compared and determined.
(Method A) A method of determining how much a feature amount of a child to be authenticated is significantly different from a feature amount reference value of an adult male. For example, assuming that the child feature amount is smaller than the adult male feature amount reference value in 7 out of 9 items extracted from the above 16 items of a to k, m, n, p and α and β, The matching ratio with the feature amount reference value is (9−7) /9×100=22.2 (%). Here, an arbitrary threshold value t1 (%) is set in advance, and if the above matching ratio 22.2 (%) is equal to or greater than the threshold value t1 (%), the authentication unit 7 sets the face data Is determined to be an adult, and the fact is notified to the authentication result section 8. If the above matching ratio 22.2 (%) is less than the threshold value t1 (%), it is determined that the child is a child, and the fact is notified to the authentication result unit 8.
[0027]
(Method B) A method of determining how much the absolute value difference between the child's characteristic amount and the adult male characteristic amount reference value is. For example, if a certain standard deviation is given to the feature value reference value of adult males, almost all adult males are included in this group. On the other hand, as the age is lower, the value of the feature value is located farther from this group, that is, in a group that should be called a child feature value distribution. Therefore, it is verified for each item how much the obtained child feature amount is different from the adult male feature amount reference value (or distribution) for each item, and the matching ratio of the feature amount reference value in the case of method A is examined. As in the case, an arbitrary threshold value s1 (mm) is provided in advance, and if the feature amount is separated by the threshold value s1 (mm) or more, the authentication unit 7 determines that the child is a child and determines that the child is a child. Tell part 8. If the distance is less than the threshold value s1 (mm), it is determined that the person is an adult, and the fact is notified to the authentication result unit 8.
[0028]
FIG. 7 is a flowchart showing a process of authentication from input of a face image by a person authentication device to determination of an adult or child in the second embodiment of the present invention. That is, FIG. 7 determines whether the image of the person to be authenticated is an adult or a child based on the information on the feature amounts a to k, m, n, p and α and β of the face data described above. It is a flowchart which shows a procedure.
[0029]
In FIG. 7, first, feature amount data is called from the data table of the feature amount calculation image processing unit 5 (step S1). Next, it is determined whether or not the called feature amount data has a beard (step S2). Here, if there is no beard (No in step S2), it is determined whether or not the called feature amount data includes a throat (step S3). Here, if there is no buddha (No in step S3), as shown in FIG. 6, the angle β of the face contour, the eye sizes c and d, the nose length g, and the mouth sizes e and f Then, the characteristic amount data such as is measured (step S4). Then, the measured feature amount data is compared with the feature amount reference value stored in the feature amount recording unit 6 (step S5), and if the measured feature amount data is less than the feature amount reference value, the eyebrows, eyes, and nose Then, the position and distance of the mouth are measured (step S6). Further, the feature amount data of the position and the distance of the part measured at this time are compared with a feature amount reference value (step S7), and if the feature amount data is less than the feature amount reference value, the authentication unit 7 performs the measurement. The face data of the image of the person is authenticated (determined) as a child, and information to that effect is notified to the authentication result unit 8 (step S8). Further, the above-mentioned determination methods of the method A and the method B are included in the step S7.
[0030]
On the other hand, in step S2, there is a beard (in the case of Yes in step S2), in step S3, there is a throat buddha (in the case of Yes in step S3), and in step S5, the angle of the face contour, the size of eyes, nose, and mouth. If is greater than or equal to the feature amount reference value, and if the feature amount data of the positions and distances of the eyebrows, eyes, nose, and mouth is greater than or equal to the feature amount reference value in step S7, the authentication unit 7 The face data is authenticated (determined) as an adult, and information to that effect is notified to the authentication result unit 8 (step S9). In this way, when it is determined that the face data of the image of the measured person is a child, a command is sent to the personal computer so that entry to a site that is not suitable for a child is restricted. Here, when the age is about 15 to 17 years, the difference between the skeleton of the adult and the child becomes small and it becomes difficult to determine. However, the authentication can be performed by appropriately changing the threshold value t (%) or the threshold value s (mm). Accuracy can be improved.
[0031]
<Third embodiment>
Next, a person authentication device according to a third embodiment of the present invention will be described. The person authentication apparatus according to the third embodiment performs person authentication when a large number of persons use a personal computer in an environment shared and used by an unspecified number of persons. PCs in an environment commonly used by an unspecified number of people, such as Internet cafes, cannot usually customize the operation level of PCs for each customer despite the fact that they are used to or unfamiliar with the operation of the user. It is considered that there is a difference in the operation level of a personal computer between a woman generally considered to be vulnerable to a machine and a man considered to be a machine expert. Therefore, it would be very convenient if the personal computer operation could be customized for each user.
[0032]
Also in this case, using the same algorithm as that of the first embodiment, the face data of the person image P captured by the image data input unit 2 of the person authentication device 1 shown in FIG.・ Extract the shape and position of the mouth. Here, in general, the skeleton of an adult female is slightly smaller than that of an adult male, so that a difference appears in the contour of the face and the degree of development of the jawbone between an adult female and an adult male. Naturally, the size of the face is smaller in women, so that components such as eyes and nose are small, and the distance between components such as eyes and nose is also short.
[0033]
Specifically, feature data of adult females such as eye sizes c and d, mouth sizes e and f, and nose length g shown in FIG. 6 are compared with adult male feature value reference values. . Also, the distance h between the eyes, the distance i between the eyebrows i, the distance j between the nose and the eye, the distance k between the nose and the mouth, the distance n between the nose and the chin, the face width p, the chin angle α, the face contour angle β, etc. A similar comparison is made for. For some adult women, the result may be that some of these features are not much different from the reference values of the features of an adult male.
[0034]
Therefore, one of the following three determination methods is compared and determined. Of course, the determination may be made by combining these methods.
(Method C) A method of determining the presence or absence of makeup. In general, it is considered that an adult woman generally wears makeup when going out. Therefore, when sampling the skin color of an adult woman, it is usually different from the feature amount of the skin color of an adult man who does not make up. The sampling points are, for example, around the eyes 11 in FIG. 3, several parts of the cheeks 12, each part of the mouth 14, and the like, and color data of these sampling points has a color bias or a specific color increase. If so, judge it as an adult woman.
[0035]
For example, a specific color such as blue, brown, or black increases around the eye 11 when an adult woman wears an eye shadow or an eye line. In addition, when the cheeks 12 of an adult woman are blusher, the cheeks 12 are biased in color such as red and pink. Further, since the mouth 14 of an adult woman is often painted with lipstick, a specific color such as red increases. In particular, since the color of the blue component is a color component (pigment) not found in human skin, whether or not makeup is applied by extracting these color components, that is, whether it is an adult female or adult male Can be determined.
[0036]
(Method D) A method of determining the number of significant differences between the feature value of an adult woman and the feature value reference value of an adult man. For example, if seven of the eleven items extracted from the 16 items a to k, m, n, p and α and β described above indicate that the feature amount of an adult female is smaller than the feature amount reference value of an adult male. For example, the matching ratio with the feature amount reference value is (11−7) /11×100=36.4 (%). An arbitrary threshold value t2 (%) is set in advance, and if the matching ratio 36.4 (%) is equal to or greater than the threshold value t2 (%), the authentication unit 7 determines that it is an adult male and notifies that fact. Notify the authentication result section 8. If the coincidence ratio 36.4 (%) is less than the threshold value t2 (%), it is determined to be an adult woman, and the fact is notified to the authentication result section 8.
[0037]
(Method E) A method for determining how much the absolute value difference between the feature value of an adult woman and the feature value reference value of an adult man is. For example, if a certain standard deviation is given to the feature value reference value of adult males, almost all adult males are included in this group. On the other hand, in the case of an adult female, the value of the feature value is located at a position distant from this group, that is, in a group that should be called a feature value distribution of the adult female. Therefore, it is verified for each item how much the feature amount obtained from the adult woman is far from the feature amount reference value (or distribution) of the adult man. That is, as in the case of comparing and judging an adult and a child in the second embodiment, an arbitrary threshold value s2 (mm) is provided in advance, and the feature amount is separated from the threshold value s2 (mm) or more. If so, it is determined to be an adult woman and that fact is communicated to the authentication result section 8. If the feature value is less than the threshold value s2 (mm), it is determined to be an adult man and the fact is notified to the authentication result unit 8.
[0038]
In other cases, it may be acceptable to judge an adult woman based on other criteria. For example, when earrings are used as earrings larger than earrings, it may be determined to be an adult woman. On the other hand, when the person has a throat and a beard, or when the hairstyle is short (for example, shaved head, square cut, sports cut, skin head, etc.), it may be determined to be an adult male.
[0039]
FIG. 8 is a flowchart illustrating a process of authentication from input of a face image by the person authentication device to determination of a woman or a man in the third embodiment of the present invention. That is, in FIG. 8, it is determined whether the image of the person to be authenticated is a woman or a man based on the information of the feature amounts a to k, m, n, p and α and β of the face data described above. It is a flowchart which shows a procedure.
[0040]
In FIG. 8, first, feature amount data is called from the data table of the feature amount calculation image processing unit 5 (step S11). Next, it is determined whether or not the called feature amount data has a beard (step S12). Here, if there is no beard (No in step S12), it is determined whether or not the called feature amount data includes a throat (step S13). Here, if there is no throat buddha (No in step S13), as shown in FIG. 6, the angle β of the face contour, the eye sizes c and d, the nose length g, and the mouth sizes e and f Are measured (step S14). Then, the measured feature amount data is compared with a feature amount reference value (step S15). If the measured feature amount data is less than the feature amount reference value, the positions and distances of eyebrows, eyes, nose, and mouth are measured ( Step S16). Further, the characteristic amount data of the position and the distance of the part measured at this time are compared with a characteristic amount reference value (step S17). If the characteristic amount data is less than the characteristic amount reference value, the authentication unit 7 performs the measurement. The face data of the image of the person is authenticated as a woman, and the information to that effect is notified to the authentication result section 8 (step S18). Further, the above-described determination methods from the method C to the method E are included in the above-described step S17.
[0041]
On the other hand, in step S12, there is a beard (in the case of Yes in step S12), in step S13, there is a throat buddha (in the case of Yes in step S13), and in step S15, the angle of the face contour, the size of eyes, nose, and mouth. Is greater than or equal to the feature amount reference value, and if the feature amount data of the positions and distances of the eyebrows, eyes, nose, and mouth is greater than or equal to the feature amount reference value in step S17, the authentication unit 7 The face data is authenticated as a male, and the information to that effect is notified to the authentication result section 8 (step S19). When a woman is thus determined, a command is sent to the personal computer so that the help function of the personal computer is displayed more carefully and easily than usual, and a voice operation guide is provided.
[0042]
<Fourth embodiment>
Next, a person authentication device according to a fourth embodiment of the present invention will be described. The person authentication apparatus according to the fourth embodiment performs person authentication when an elderly person uses a personal computer. That is, if it is convenient to be able to customize the operation of the personal computer for each user, there is a case where an elderly person is an operation target, in particular. Also in the case of the fourth embodiment, using the same algorithm as that of the first embodiment, the face data of the person image P captured by the image data input unit 2 of the person authentication device 1 shown in FIG. , Shapes and positions of eyebrows, eyes, nose, mouth, etc. are extracted. In addition, the degree of sagging of the cheeks and the color and luster / stain / wrinkles of the skin are extracted as elements for determining the face of the elderly.
[0043]
At this time, since it is difficult to numerically compare the degree of sagging of the cheeks and the stains, etc., streaks (wrinkles) of the forehead, the outer corner of the eyes, the cheeks, and the mouth are detected from the sampling result of the face image. Then, the streak (wrinkle) is defined as a set of continuous points equal to or greater than v (mm), and the number is compared with a feature amount reference value of an adult male. Here, a threshold value w (book) is set in advance, and if the streak (wrinkle) of the sampling result of the face image is equal to or greater than the threshold value w (book), the authentication unit 7 determines that the person image is an elderly person. And notifies the authentication result unit 8 of the determination. If the streak (wrinkle) is less than the threshold value w (book), the image of the person is determined not to be an elderly person, and the fact is transmitted to the authentication result unit 8. Note that v and w are arbitrary rational numbers.
[0044]
FIG. 9 is a flowchart showing a process of authentication from input of a face image by a person authentication device to determination of an elderly person in the fourth embodiment of the present invention. In FIG. 9, first, feature amount data is called from the data table of the feature amount calculation image processing unit 5 (step S21). Then, the wrists, the corners of the eyes, the nose, the corners of the mouth and the cheeks are measured from the feature amount data (step S22). Next, the measured feature value data is compared with the feature value reference value (step S23). If the measured feature value data (the outer corner of the eye, the nostrils, the corners of the mouth, the wrinkles of the cheeks, etc.) is equal to or more than the feature value reference value, Then, the color and luster of the forehead and cheek are measured (step S24). Further, the feature amount data of the color and gloss of the forehead and cheek measured at this time is compared with a feature amount reference value (step S25), and if these feature amount data are less than the feature amount reference value (for example, forehead or cheek) If the color or luster of the image is different from the natural color), the authentication unit 7 authenticates that the measured face data of the image of the person is an elderly person and notifies the authentication result unit 8 of the information (step S26). ).
[0045]
On the other hand, in step S23, the feature amount data of the corners of the eyes, the nostrils, the corners of the mouth, and the wrinkles of the cheeks are smaller than the feature amount reference values, or in step S25, the feature amount data of the forehead, cheek color and gloss measured in step S24 are used. When the feature value is equal to or larger than the reference value (that is, the forehead and cheeks are close to the natural color and luster), the authentication unit 7 authenticates the image of the measured person as a non-elderly person, and outputs information indicating that to the effect. The notification is made to the unit 8 (step S27). If the person is judged to be an elderly person in this way, the help function of the personal computer is displayed more carefully and clearly than usual, the displayed characters are larger than usual, and the volume of the operation guide by voice is increased. Or send a command to the personal computer.
[0046]
<Fifth embodiment>
Next, a person authentication device according to a fifth embodiment of the present invention will be described. The person authentication apparatus according to the fifth embodiment performs person authentication at regular intervals for a person who uses a personal computer. For example, when sharing a personal computer at home with the whole family, once the computer is authenticated with the father at the start of use, even if the child has been using the computer on his way, the child has been using it for a long time. However, the current user is recognized by the personal computer as a father (adult). In order to solve this, verification of face data of a person is repeated at regular intervals, and authentication is performed each time. This eliminates the inconvenience that occurs when a person who has performed person authentication at the time of activation is replaced with another person.
[0047]
<Sixth Embodiment>
Next, a person authentication device according to a sixth embodiment of the present invention will be described. The person authentication device according to the sixth embodiment performs person authentication so that a personal computer used by an unspecified number of people is not used for a crime. For example, a personal computer in an environment that is shared and used by an unspecified number of people, such as an Internet cafe, has a risk of being used for a crime by utilizing the fact that the user is unspecified. For example, using anonymity at the time of operation to send unsolicited e-mail equivalent to stalking, or fraudulent activities in Internet transactions such as securities have actually occurred.
[0048]
Therefore, by applying the function of the person authentication device of the present invention, the image of the face captured by the image data input unit at the start of use of the personal computer is registered and stored as image data (face data). This image data is normally closed from the viewpoint of privacy protection, but it is assumed that it can be used after a certain procedure such as the law. Thereby, for example, when a spam mail is received, if the source Internet cafe is specified from the IP address or the like, the criminal can be specified from the image data of the face image registered and stored inside. . In addition, if the image data of the wanted man is registered in the personal computer, the image data of the wanted man registered and accumulated at the start of using the personal computer can be collated. Here, when it is determined that they match, it is also possible to notify the police. Thus, it is also effective in criminal investigations.
[0049]
The embodiment described above is an example for describing the present invention, and the present invention is not limited to the above embodiment, and various modifications are possible within the scope of the invention. For example, in the above-described embodiment, the case where the personal authentication device is applied to the personal computer has been described. However, the present invention is not limited to this. For example, when the personal authentication device is used at a financial institution or when a credit is used at a department store or the like. It can also be applied to authentication. That is, the personal authentication device of the present invention can be widely applied to a case where a desired device is operated, a case where it is determined whether or not an operation right is given to the device, or a case where authentication is performed.
[0050]
By the way, as is clear from the description of the person authentication device described in each of the above embodiments, the authentication unit is configured by the feature amount data acquired from the feature amount calculation image processing unit and the feature amount reference value acquired from the feature amount recording unit. Comparing and analyzing the data and identifying the person in broad categories such as old, young, male, and female based on the analysis result, and determining the applicable range of the desired device according to the identified person This is a preferred embodiment of the present invention. With this, for example, when an elderly person or a woman who is generally said to be vulnerable to the operation of a personal computer uses a personal computer, the help function is displayed more carefully and easily than usual, and a voice operation guide is provided. , It is possible to set the environment in which the PC can be easily used.
[0051]
Also, a person authentication device that captures an image of a person who intends to perform an operation of a desired device by an imaging unit, extracts a face image from the image of the person, and authenticates the person, starts using the desired device. Sometimes a face image of a person is captured by the image data input means and analyzed as face image data, and the face image data is compared with previously registered and stored face image data to determine whether they match. It is a preferred embodiment of the present invention to configure such that if they match, a right such as file access or Internet connection by a desired device (for example, a personal computer) is permitted. As a result, it is possible to prevent a person who is not authorized to operate the personal computer from illegally accessing a file or a child from accessing an undesirable site.
[0052]
Also, it is determined whether the image data of the face of the person who wants to perform the operation of the desired device matches the image data of the face registered and stored in advance to determine whether or not the image data matches. It is a preferable aspect of the present invention to configure so as to partially limit the applicable range of (for example, a personal computer). Thus, for example, when sharing a personal computer in the same family or workplace, or when operated by a person other than the main registrant, it can be used within a range where confidential contents cannot be disclosed.
[0053]
Also, it is determined whether the image data of the face of the person who is to perform the operation of the desired device matches the image data of the face registered and stored in advance to determine whether or not they match. It is a preferred embodiment of the present invention to configure so as to partially restrict file access by a device (for example, a personal computer) and Internet connection. Thus, for example, when sharing a personal computer with the same family, it is possible to restrict access to some undesirable sites only when a child operates the personal computer.
[0054]
Also, image data of a face of a person who intends to perform an operation of a desired device is fetched and analyzed, and if the analysis result does not satisfy a certain condition, a part of a possible range of the desired device (for example, a personal computer) is reduced. Restrictive configuration is a preferred aspect of the present invention. Thus, for example, when sharing a personal computer with the same family, if a person other than the main registrant, such as a child, operates the personal computer, general use is possible, but careless access to important files is inadvertent. Content cannot be erased or rewritten.
[0055]
In addition, image data of a face of a person who intends to perform an operation of a desired device is captured and analyzed. If the analysis result does not satisfy a certain condition, a file access by a desired device (for example, a personal computer) or an Internet connection is performed. Is a preferred embodiment of the present invention. For example, when sharing a personal computer at the same workplace, when a person other than the main registrant operates, or when a person who does not satisfy the conditions above the specified responsibilities tries to operate, general use However, confidential contents such as personnel records cannot be disclosed.
[0056]
【The invention's effect】
As described above, according to the person authentication device of the present invention, by extracting features from a face image of a person who intends to operate a personal computer, it is possible to restrict the operation of the personal computer by a person who does not have the operation right. . For example, when accessing a file or site that needs to be restricted for security or confidentiality, perform personal authentication when accessing a file or site, or access a site that is not suitable for children to watch in an environment where a personal computer is shared at home. This can be used when performing person authentication when restricting. Further, such a person authentication device is also used for a personal computer in an environment where an unspecified number of people share and use it, such as an Internet cafe. For example, in a personal computer in an environment in which an unspecified number of people share and use, the age and gender of the person are determined, and the present invention is also applied to a case where the operation assistance suitable for each user is performed. This can be easily realized by performing person authentication using a person authentication device. That is, by using the person authentication device of the present invention, security, convenience, and operability when operating the personal computer can be further improved.
[Brief description of the drawings]
FIG. 1 is a block diagram of a person authentication apparatus according to the present invention which fetches an image of a person's face by an image data input unit and compares the image with a registered face image to authenticate a person;
FIG. 2 is an image diagram of a face image extracted by a face area extracting unit 3 in FIG. 1;
FIG. 3 is an image diagram for extracting a feature amount of face data from the face image shown in FIG. 2;
FIG. 4 is an image diagram for extracting a feature amount of face data from a face image captured from an image of an adult person in the person authentication device shown in FIG. 1;
FIG. 5 is an image diagram for extracting a feature amount of face data from an image of a face captured from an image of a child person in the person authentication device shown in FIG. 1;
FIGS. 6A and 6B are conceptual diagrams in a case where a feature amount of face data is extracted in the person authentication apparatus shown in FIG. 1, wherein FIG. 6A shows extraction of a part of the face data, and FIG. 5 shows the extraction of the feature amount of another part of FIG.
FIG. 7 is a flowchart illustrating a process of authentication from input of a face image by a person authentication device to determination of an adult or child in the second embodiment of the present invention.
FIG. 8 is a flowchart showing a process of authentication from input of a face image by a person authentication device to determination of a woman or a man in the third embodiment of the present invention.
FIG. 9 is a flowchart illustrating a process of authentication from input of a face image by a person authentication device to determination of an elderly person in a fourth embodiment of the present invention.
[Explanation of symbols]
1 person authentication device
2 Image data input section
3 face area extraction unit (face area extraction means)
4 feature value detection unit (feature value detection means)
5. Feature calculation image processing unit (feature calculation image processing means)
6. Feature recording unit (feature recording means)
7 Authentication section (authentication means)
8 Authentication result section
9 Control unit
10 eyebrows
Eleventh
12 cheeks
13 Facial contours and jaws
14 mouths
P People image
a Length of eyebrows (between eyebrows and eyebrows)
b Eyebrow height
c Width of eyes (from inner to outer corners)
d Eye width
e Width of mouth
f Mouth height
g Nose length
h Spacing between eyes (left eye to right eye)
i Interval between both eyebrows (left eyebrow to right eyebrow)
j Distance between nose and eyes (nose-to-eye)
k Distance between nose and mouth (nose-to-mouth corner)
m Distance between nose and eyebrow (nose to eyebrow)
n Distance between nose and chin
p face width
α Angle between chin and jaw base (jaw angle)
β Face contour angle

Claims (2)

The image of the face of the person is extracted from the image of the person obtained by imaging the person who is going to perform the operation of the desired device, and the authentication of the person and the instruction of the applicable range to the device are performed. A person authentication device,
Face area extraction means for capturing the image of the person and extracting an image area of the face,
A feature amount detecting unit that detects, as a feature amount, sampling data for each predetermined portion of the face image region extracted by the face region extracting unit;
A feature amount calculation image processing unit that calculates the feature amount detected by the feature amount detection unit to create feature amount data;
Feature amount recording means for recording feature amount reference value data previously obtained from feature amount data of a plurality of persons;
Comparing and analyzing the feature amount data obtained from the feature amount calculation image processing unit and the feature amount reference value data obtained from the feature amount recording unit,
Identifying the attribute of the person based on the analysis result,
Setting a possible range of the desired device according to the specified attribute,
An authentication unit that instructs the device to indicate the applicable range,
A person authentication device provided. The face area extracting means extracts the image area of the face by capturing the image of the person every predetermined time or every time a predetermined change occurs in the attribute of the person,
2. The apparatus according to claim 1, wherein the authentication unit compares and analyzes the feature amount data and the feature amount reference value data each time the face area detection unit extracts a face image area. 3. .

Images