JP4585254B2 - Authentication system - Google Patents

Description

  The present invention relates to authentication, specifically, an authentication method, apparatus, and program for performing authentication using a face image at the time of authentication of a person to be authenticated and a face image registered at a predetermined age of the person. .

  Conventionally, a face image at the time of person authentication (hereinafter referred to as an “age face image”) and a face image registered at a predetermined age of the person (hereinafter referred to as a “registered face image”) are collated for pattern matching or the like. Realization is performed (for example, refer nonpatent literature 1). However, since the human face has different skin conditions such as strength and weakness of the wrinkle component with aging, it is registered in advance with the current face image of the target person as described in Non-Patent Document 1. In a system that performs authentication by pattern matching with a person's face image, there is a problem that the accuracy of the authentication is lost when the number of years has passed since the registration of the face image. Therefore, it is conceivable to update the registration of facial images at an appropriate time, but the registration update work becomes a huge burden when the number of registered people increases, and the system aims to find criminals. In some cases, registration update work is often impossible, which is unrealistic.

  Therefore, in Patent Document 1, a face image input space at the time of authentication is projected onto a different space, and a change over time is canceled by canceling a pattern change unnecessary for authentication, or a change in illumination conditions at the time of capturing a face image. A system has been proposed that reduces the effects of authentication and improves authentication accuracy.

  Among them, as a method for canceling the wrinkle component and the like over time, for example, it is conceivable to apply a process of projecting the face image input space at the time of authentication to the frequency space and suppressing the wrinkle component in a high frequency band. The method described in Document 2 can be used.

  The method described in Patent Document 2 utilizes the fact that many components such as wrinkles and spots (hereinafter generally referred to as wrinkle components) exist as small-amplitude signals in the high-frequency components of the image, thereby reducing the small amplitude in the image. The ε-filter (ε-separated non-linear digital filter) devised to separate and suppress the high-frequency noise component is applied to the removal of wrinkles and spots. Since the ε-filter has the property of flattening only small amplitude level changes in the image signal, the image processed using the ε-filter preserves edges with steep level changes, and The sharpness of the is almost not impaired.

  The ε-filter basically serves to subtract a value obtained by applying a nonlinear function to the amount of change in amplitude level from the original image signal. This nonlinear function is a function that sets the output to 0 when the amplitude of the signal is larger than a predetermined threshold. The output of this nonlinear function is considered to correspond to the wrinkle component in the image. That is, when the ε-filter is applied, the output of the nonlinear function is 0 at a portion in the image where the amplitude is larger than the above-mentioned threshold value. In the processed image, the original signal of the part is retained. On the other hand, in the part where the amplitude is equal to or smaller than the threshold value, in the processed image, the signal value of the part is output from the original signal value as a nonlinear function. (The absolute value is greater than 0). By doing this, although it is not so-called noise, in the wrinkles and spots such as wrinkles that show small amplitude light and dark changes, the light and dark changes are smoothed, and wrinkles and spots can be made inconspicuous, and edges with large amplitude Can hold the part.

The processing for suppressing the wrinkle component described in Non-Patent Document 2 is applied to the authentication system of Patent Document 1, and the image obtained by suppressing the wrinkle component on the face image at the time of authentication is collated with the registered face image. By performing authentication, it is possible to reduce the influence of an increase in wrinkle components due to the passage of time, and it is possible to perform accurate authentication.
JP 2000-30065 A "Face recognition system using facial images", IEICE Technical Report PRMU 97-50, June 1997, by Yamaguchi et al. Arakawa et al., "Color face image processing with vector ε-filter-Removal of flaw components" March 1998 Proceedings of the IEICE General Conference, D-11-43, PP143-

  However, in the case of a registered face image registered after a predetermined age, for example, 25 years old, the registered face image also includes wrinkle components or noise components. On the other hand, in the method described in Patent Document 1 in which processing for canceling an unnecessary pattern such as suppression of wrinkle components is performed on the current face image at the time of authentication and collation with the registered face image is performed, the wrinkle component of the current face image Since it is difficult to suppress the above-mentioned mode so as to be a wrinkle mode at the time of registration, there is a problem that the accuracy of authentication cannot be maintained if the suppression level is too strong or too weak.

  Moreover, although many wrinkle components exist in the high frequency band, they exist over each frequency band from the high frequency band to the low frequency band. The wrinkle suppression processing method using the ε-filter described above extracts components such as wrinkles and stains in only one frequency band and subtracts the components extracted from the original image. There is a problem that extraction leakage of components such as the above will occur. Therefore, even if the wrinkle suppression process is performed using the wrinkle component extracted in one frequency band, there is a wrinkle component that remains as it is, and such a wrinkle suppression process method is applied to the authentication system described in Patent Document 1. Even so, there is a limit to increasing the accuracy of authentication. In addition, the above-described method of extracting components such as wrinkles in one frequency band enhances filtering in this one frequency band in order to enhance the effect of suppressing wrinkles and spots, that is, wrinkles and spots described above. Artifacts are likely to occur because there is no choice but to increase the threshold for extracting. As a result of strengthening the wrinkle suppression process in order to enhance the effect of the wrinkle suppression process, if an artifact occurs in the processed pixel, there is a problem that the accuracy of authentication is returned and lowered.

  The present invention has been made in view of the above circumstances, and when performing authentication using a face image at the time of authentication of a person to be authenticated and a registered face image of this person, the time of authentication is the time from the time of registration. An object of the present invention is to provide an authentication method and apparatus and a program therefor that can prevent a decrease in the accuracy of authentication even after passing.

The authentication method of the present invention is an authentication method for performing authentication using an actual face image that is a face image at the time of authentication of a person to be authenticated, and a registered face image registered at a predetermined age of the person.
Each of the current age face image and the registered face image is subjected to a wrinkle component suppression process to create a corresponding authentication image,
The authentication is performed by collating the two authentication images.

  In addition, it is preferable that the wrinkle component suppression process is performed on the current age face image with a suppression strength stronger than the suppression strength of the wrinkle component suppression processing on the registered face image.

In the authentication method of the present invention, the wrinkle component suppression processing includes a plurality of band-limited images representing components for each of a plurality of frequency bands of the target image based on the target image to be the current age face image or the registered face image. Create
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to get the pixel values of multiple converted images,
Multiplying a pixel value of the plurality of converted images by a predetermined adjustment coefficient to obtain a pixel value of the adjustment component image;
It is preferable that the pixel value of the adjustment component image be subtracted from the pixel value of the target image.

  In this case, as the adjustment coefficient, the same value may be used for each pixel of the converted image, but the target image (current face image or registered face image) used when creating the converted image is used. It is preferable to use a pixel determined according to the pixel value.

An authentication apparatus according to the present invention is an authentication apparatus that performs authentication using an actual face image, which is a face image at the time of authentication of a person to be authenticated, and a registered face image registered at a predetermined age of the person. ,
A wrinkle component suppression unit that performs a wrinkle component suppression process on the current face image and the registered face image, respectively, and creates corresponding authentication images,
And a verification unit that performs the authentication by comparing the two authentication images obtained by the wrinkle component suppression unit.

  It is preferable that the wrinkle component suppression unit performs the wrinkle component suppression process on the current face image with a suppression strength stronger than the suppression strength of the wrinkle component suppression process on the registered face image.

Further, the wrinkle component suppression means creates a plurality of band limited images representing components for a plurality of frequency bands of the target image based on the target image to be the current face image or the registered face image,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to get the pixel values of multiple converted images,
Multiplying a pixel value of the plurality of converted images by a predetermined adjustment coefficient to obtain a pixel value of the adjustment component image;
It is further preferable to perform a process of subtracting the pixel value of the adjustment component image from the pixel value of the target image.

  More preferably, the adjustment coefficient is determined according to a pixel value of the target image.

  Moreover, you may provide the authentication method of this invention as a program for making a computer perform the process of the authentication method of this invention.

  According to the authentication method and apparatus of the present invention, the wrinkle component suppression process is performed on both the current face image at the time of authentication of the person to be authenticated and the registered face image of the person, and the wrinkle component is suppressed. Since the authentication is performed by using the authentication face image as the authentication face image and the registered face image in which the wrinkle component is suppressed, the wrinkle component is suppressed or removed from the authentication image. Therefore, the influence of the wrinkle component on the authentication accuracy can be prevented, and the authentication accuracy can be maintained.

  In addition, the appearance of wrinkles in the face image varies depending on the illumination conditions. Therefore, according to the authentication method and apparatus of the present invention, it works to suppress or remove the wrinkle component. It is also possible to reduce the influence due to the difference in illumination conditions during imaging.

  In addition, since the age at the time of authentication is higher than the age at the time of registration, the wrinkle component in the current-age face image is larger than the wrinkle component in the registered face image, and thus the suppression strength of the wrinkle component suppression process for the registered face image If the wrinkle component suppression process is performed on the current-age face image with a strong suppression strength, the accuracy of authentication can be further increased.

  In addition, as the wrinkle component suppression processing, first, a plurality of band-limited images representing components for each of a plurality of different frequency bands of the present-age face image created based on the target image (current-age face image or registered face image) Is subjected to nonlinear conversion processing to obtain a plurality of converted images. This non-linear conversion process is a process of making the absolute value of the output value less than or equal to the absolute value of the input value, and for an input value whose absolute value is less than or equal to a predetermined threshold value, the larger the absolute value of the input value, While the absolute value of the value is increased, for an input value whose absolute value is greater than a predetermined threshold, the absolute value of the output value is equal to or less than the absolute value of the output value corresponding to the predetermined threshold. The converted image obtained by this nonlinear conversion processing represents a component having a small amplitude such as wrinkles, spots or noises in the frequency band corresponding to the converted image. The present invention suppresses the wrinkle component by subtracting the pixel value of each adjustment component image obtained by multiplying the pixel values of the plurality of converted images by a predetermined adjustment coefficient from the pixel value of the original target image. Since the pixel value of the authentication image is obtained, it is possible to effectively suppress components such as wrinkles, spots, and noises in a plurality of frequency bands of the target image, thereby further maintaining the accuracy of authentication. .

  Further, in the conventional technique for extracting components such as wrinkles and spots in one frequency band, in order to extract the wrinkle components effectively, non-linear processing in this one frequency band is strengthened (that is, wrinkles, spots, etc.). The threshold for extracting the component needs to be increased), and artifacts are generated in the processed image (in the present invention, the authentication image) in which the wrinkle component is suppressed, leading to deterioration in image quality or authentication accuracy. There was a problem. On the other hand, in the present invention, components such as wrinkles, spots or noises are extracted and suppressed from a plurality of frequency bands, and a wrinkle component suppressing effect that does not require so much enhancement of nonlinear processing in each frequency band is achieved. Therefore, the generation of artifacts can be prevented, and an authentication image with high image quality can be obtained.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration of an authentication apparatus and an authentication apparatus according to an embodiment of the present invention and a program therefor. Note that the authentication device of the present embodiment is realized by executing an authentication processing program read into the auxiliary storage device on a computer (for example, a personal computer). The authentication processing program is stored in an information storage medium such as a CD-ROM or distributed via a network such as the Internet and installed in a computer.

  Further, since the image data represents an image, the following description will be given without particularly distinguishing the image from the image data.

  As illustrated in FIG. 1, the authentication apparatus according to the present embodiment includes an imaging unit 1, an authentication unit 100, a first database 120, and a second database 140. The face image D0 is obtained by capturing an image of the target person, and the first database 120 stores the registered face image Dg of each person in association with the age at the time of registration. The database 140 of FIG. 2 stores various parameters provided to the authentication unit 100. The authentication unit 100 stores the face image D0, the registered face image Dg stored in the first database 120, and the second image. Authentication is performed using parameters stored in the database 140.

  FIG. 2 is a block diagram showing a configuration of the authentication unit 100 in the authentication apparatus shown in FIG. 1. As shown in the figure, the authentication unit 100 includes an input unit 2, an image generation unit 60, and a verification unit 70. Thus, the input unit 2 is for causing the person to be authenticated to input information that can identify the person (for example, the password P given to the person). The face image D0 obtained by the imaging unit 1 is subjected to processing for suppressing wrinkles or noise components (hereinafter generally referred to as wrinkle components) to generate a first authentication image D1, and is stored in the first database 120. The registered face image Dg is processed to suppress the wrinkle component to generate the second authentication image D2. The verification unit 70 generates the first authentication image generated by the image generation unit 60. Image D1 and second authentication In which collates the image D2.

  3 is a block diagram illustrating a configuration of the image generation unit 60 in the authentication unit 100 illustrated in FIG. 2. As illustrated, the image generation unit 60 includes the target specifying unit 3, the YCC conversion unit 5, and the blurred image. The image forming apparatus includes a creating unit 10, a band limited image creating unit 20, a wrinkle component extracting unit 30 for extracting wrinkles, an authentication image creating unit 40, and a combining unit 50.

  The blurred image creating means 10 creates a plurality of blurred images S1, S2,... Sn (n: integer greater than or equal to 2) of the original image S0 having different frequency response characteristics. Is to create a plurality of band limited images T1, T2,... Tn using the original image S0 and the blurred images S1, S2,. , T2,... Tn are each subjected to nonlinear transformation processing to extract wrinkle components Q1, Q2,... Qn in the frequency band corresponding to each band-limited image. Is used to suppress the wrinkle component in the original image S0 using the original image S0 and the wrinkle components Q1, Q2,. Since these means perform processing in the luminance space, the YCC conversion means 5 performs YCC conversion on the face image D0 (R0, G0, B0) obtained by the imaging unit 1 to generate the face image D0. The luminance Y0 (the luminance component Y0 constitutes the above-described original image S0 corresponding to the face image D0), the color differences Cb0, Cr0, and the registered face image Dg (Rg, Rg, stored in the first database 120). Gg, Bg) is subjected to YCC conversion to obtain luminance Y0 of the registered face image Dg (the above-described original image S0 corresponding to the registered face image Dg is constituted by these luminance components Y0), and color differences Cbg, Crg. The synthesizing unit 50 uses the pixel value Y1 of the first authentication image S′1 corresponding to the face image D0 obtained by the authentication image creating unit 40 and the YCC conversion unit 5. A first authentication image D1 (Y1, Cb0, Cr0) is obtained by synthesizing an image composed of the obtained color differences Cb0 and Cr0, and a registered face image Dg obtained by the authentication image creation means 40 is obtained. Are combined with the pixel value Y1 of the second authentication image S′2 corresponding to the color difference Cbg and Crg obtained by the YCC conversion means 5, and the second authentication image D2 (Y1, Y2) is synthesized. Cbg, Crg). Here, the detail of each structure of the image generation part 60 is demonstrated.

  The YCC conversion means 5 converts the R, G, and B values of the face image D0 and the registered face image Dg into a luminance value Y and color difference values Cb and Cr according to the following equation (1).

Y = 0.2990 × R + 0.5870 × G + 0.1140 × B
Cb = −0.1687 × R−0.3313 × G + 0.5000 × B + 128 (1)
Cr = 0.5000 × R−0.4187 × G−0.0813 × B + 128

The blur image creation unit 10 creates a plurality of blur images using the luminance value Y0 obtained by the YCC conversion unit 5. FIG. 4 is a block diagram illustrating a configuration of the blurred image creation unit 10. As shown in the figure, the blurred image creation means 10 includes a filtering means 12 that performs filtering processing to obtain filtered images B1, B2,... Bn, an interpolation means 14 that performs interpolation processing on each filtering image, and filtering. The control means 16 which controls the means 12 and the interpolation means 14 is provided. The filtering means 12 performs a filtering process using a low-pass filter. As this low-pass filter, for example, a filter F substantially corresponding to a 5 × 1 grid-like one-dimensional Gaussian distribution as shown in FIG. 5 is used. be able to. This filter F has σ = 1 in the following equation (2).

  The filtering unit 12 performs a filtering process on the entire image to be processed by performing a filtering process on the image to be processed in the x direction and the y direction of the pixel using such a filter F.

FIG. 6 shows details of processing that the control unit 16 of the blurred image creation unit 10 causes the filtering unit 12 and the interpolation unit 14 to perform. As shown in the figure, the filtering unit 12 first performs a filtering process on the original image S0 (Y0) every other pixel using the filter F shown in FIG. A filtering image B1 (Y1) is obtained by this filtering process. The size of the filtering image B1 is 1/4 of the size of the original image S0 (1/2 in the x direction and y direction, respectively). Next, the filtering unit 12 performs the filtering process by the filter F every other pixel on the filtering image B1 (Y1) to obtain the filtering image B2 (Y2). The filtering means 12 repeats the filtering process by such a filter F, and obtains n filtered images Bk (k = 1 to n). The size of the filtering image Bk has become a ^{1/2 2K} of the size of the original image S0. FIG. 7 shows the frequency characteristics of each filtered image Bk obtained by the filtering means 12 when n = 3 as an example. As shown in the figure, the response of the filtered image Bk is such that the higher the k, the higher the frequency component is removed.

  In the authentication apparatus according to the present embodiment, the filtering unit 12 performs the filtering process on the x direction and the y direction of the image by the filter F shown in FIG. 5, but the 5 × as shown in FIG. The filtering process may be performed on the original image S0 and the filtering image Bk at once by the two-dimensional filter 5.

The interpolation unit 14 performs an interpolation process on each filtering image Bk obtained by the filtering unit 12 so that the size of each filtering image Bk is the same as that of the original image S0. There are various interpolation processing methods such as a B-spline method. In the first embodiment, the filtering unit 12 uses a filter F based on a Gaussian signal as a low-pass filter. A Gaussian signal is also used as an interpolation coefficient for performing an interpolation operation of the interpolation process, and this interpolation coefficient is approximated to σ = 2 ^K−1 in the following equation (3).

When interpolating the filtering image B1, since k = 1, σ = 1. The filter for performing interpolation when σ = 1 in the above equation (3) is a 5 × 1 one-dimensional filter F1 as shown in FIG. The interpolating means 14 first enlarges the filtered image B1 to the same size as the original image S0 by interpolating pixels with a value of 0 every other pixel to the filtered image B1, and then enlarges the filtered image B1 to the same size as the original image S0. Then, a filtering process is performed by the filter F1 shown in FIG. 9 to obtain a blurred image S1. The blurred image S1 has the same number of pixels as the original image S0, that is, the same size as the original image S0.

  Here, the filter F1 shown in FIG. 9 is a 5 × 1 filter. However, before applying the filter F1, pixels having a value of 0 are interpolated with respect to the filtered image B1 every other pixel. The interpolation processing by the means 14 is substantially performed by two types of filters: a 2 × 1 filter (0.5, 0.5) and a 3 × 1 filter (0.1, 0.8, 0.1). Equivalent to filtering.

  When the interpolation means 14 performs interpolation on the filtered image B2, k = 2, so σ = 2. In the above equation (3), the filter corresponding to σ = 2 is an 11 × 1 one-dimensional filter F2 shown in FIG. The interpolation unit 14 first expands the filtering image B2 to the same size as the original image S0 by interpolating three pixels each having a value of 0 every other pixel with respect to the filtering image B2, and then expanding the image Is subjected to a filtering process by a filter F2 shown in FIG. 10 to obtain a blurred image S2. The blurred image S2 has the same number of pixels as the original image S0, that is, the same size as the original image S0.

  Similarly, the filter F2 shown in FIG. 10 is an 11 × 1 filter, but before applying the filter F2, three pixels each having a value of 0 are interpolated with respect to the filtered image B2 by three. Therefore, the interpolation processing by the interpolation means 14 is substantially performed by a 2 × 1 filter (0.5, 0.5) and a 3 × 1 filter (0.3, 0.65, 0.05), ( This is equivalent to filter processing using four types of filters (0.3, 0.74, 0.13) and (0.05, 0.65, 0.3).

In this way, the interpolation means 14 interpolates (2 ^K −1) pixels each having a value of 0 every other pixel with respect to each filtering image Bk, so that the filtering image Bk is the same as the original image S0. For a filtered image Bk that is enlarged to size and interpolated with pixels having a value of 0, the length created based on the above equation (3) is (3 × 2 ^K −1). A blurred image Sk is obtained by performing a filtering process.

  FIG. 11 shows the frequency characteristics of each blurred image Sk obtained by the blurred image creation means 10 when n = 3 as an example. As shown in the figure, the blurred image Sk is such that the higher the k, the higher frequency components of the original image S0 are removed.

  The band-limited image creating means 20 uses the blurred images S1, S2,... Sn obtained by the blurred image creating means 10 and components for each of a plurality of frequency bands of the original image S0 according to the following equation (4). .., Tn are generated.

Tm = S (m−1) −Sm (4)
Where m is an integer between 1 and n

FIG. 12 shows the frequency characteristics of each band limited image Tm obtained by the band limited image creating means 20 when n = 3 as an example. As illustrated, the band limited image Tm represents a component in the low frequency region of the original image S0 as m increases.

  The wrinkle component extraction unit 30 performs non-linear transformation on each band limited image Tm (m = 1 to n) obtained by the band limited image creating unit 20 to generate wrinkles in the frequency band corresponding to each band limited image Tm. , Spots, noise, and the like (hereinafter collectively referred to as wrinkle components) Q1, Q2,..., Qn. This non-linear transformation is a process of making the output value equal to or less than the input value, and for an input value less than or equal to a predetermined threshold value, the larger the input value, the larger the output value, but greater than the predetermined threshold value. For an input value, the output value is equal to or less than the output value corresponding to the predetermined threshold value, and is performed by a function f as shown in FIG. . In the figure, the broken line indicates a function in which the output value = the input value, that is, the slope is 1. As shown in the figure, the non-linear transformation function f used for the wrinkle component extraction means 30 has a slope of 1 when the absolute value of the input value is smaller than the first threshold Th1, and the absolute value of the input value is the first value. When the value is greater than or equal to the threshold value and less than or equal to the second threshold value Th2, the slope is smaller than 1. When the absolute value of the input value is greater than the second threshold value Th2, the output value is smaller than the absolute value of the input value. This is a function having a constant value M. The function f may be the same for each band limited image, but may be different for each band limited image.

  The wrinkle component extraction unit 30 uses the luminance value of each band limited image as an input value, performs nonlinear conversion on each band limited image using the nonlinear conversion function f shown in FIG. The configured wrinkle component Qm (m = 1 to n) in the frequency band corresponding to each band limited image is extracted and output to the authentication image creating means 40.

  Thus, the wrinkle component Qm extracted from the face image D0 and the wrinkle component Qm extracted from the registered face image Dg are input to the authentication image creating means 40, respectively.

  On the other hand, the object specifying means 3 changes the password P from the first database 120 (details of the first database 120 will be described later) based on the password P input via the input unit 2 of the authentication unit 100. The registered face image Dg stored correspondingly, the age at the time of registration of the person corresponding to the registered face image Dg, the date of registration, and the gender of the person are read out, and the registered face image Dg is read to the verification unit 70 of the authentication unit 100. On the other hand, information indicating the age at which the person corresponding to the registered face image Dg is registered (referred to as registered age), the gender of the person, and the registration date is output to the authentication image creating means 40.

  FIG. 15 shows the contents of data in the first database 120. As shown in the figure, in the first database 120, the person's name, password P, gender, registered face image Dg, age at registration, registration The days are stored in association with each other. The target specifying unit 3 reads the registered face image Dg and other data based on the password P of the person to be authenticated input via the input unit 2.

  The authentication image creation means 40 includes an original image S0 that is a luminance component of the face image D0, wrinkle components Q1, Q2,..., Qn in the original image S0 obtained by the wrinkle component extraction means 30, and second images. The first authentication image S′1 (Y1) is created using the various parameters stored in the database 140, and the original image S0 serving as the luminance component of the registered face image Dg and the wrinkle component extraction means 30 are used. A second authentication image S′2 (Y1) is created using the obtained wrinkle components Q1, Q2,..., Qn in the original image S0 and various parameters stored in the second database 140. FIG. 14 shows the configuration. As shown in the figure, the authentication image creation means 40 includes a current age calculation unit 42, a first parameter setting unit 43, a second parameter setting unit 44, a first creation execution unit 45, and a second And a creation execution unit 46. The current age calculation unit 42 calculates the current age of the person to be authenticated (hereinafter referred to as the current age) based on the registered age output from the target specifying unit 3 and the date at the time of authentication, and the first parameter This is output to the setting unit 43. In the authentication apparatus according to the present embodiment, the authentication image creating unit 40 reduces the age of the person to be authenticated from the face image D0 to, for example, 20 years old with almost no wrinkles on the face from the current age. An aged image, that is, an image in which the wrinkle component is suppressed, and an aged image in which the age of the person to be authenticated is aged from the registered age to 20 years from the registered face image Dg, that is, the wrinkle component An image in which the image is suppressed is created.

  Here, first, the process of creating the first authentication image S′1 corresponding to the face image D0, that is, the process by the first parameter setting unit 43 and the first creation execution unit 45 will be described.

  The first parameter setting unit 43 uses the current age calculated by the current age calculation unit 42 and the various parameters stored in the second database 140 to provide a first image corresponding to the face image D0 of the current age. A parameter W1 for creating the authentication image S′1 (Y1) is set. Here, the processing by the first parameter setting unit 43 will be described with reference to the second database 140 shown in FIG.

  As shown in FIG. 16, the second database 140 is further composed of three types of database A, database B, and database C.

  The database A stores the difference between the current age and 20 years old, that is, the age of decrease N1 from the current age to 20 years and the coefficient α corresponding to the age of decrease N1 in association with each other. The coefficient α increases as the age of aging N1 increases. In the illustrated example, the coefficient α is 5 years or less, 5 years to 10 years or less, 10 years to 15 years, 15 to 20 years, .., 0.1, 0.2, 0.5, 0.6,... First, the first parameter setting unit 43 first calculates the current age calculated by the current age calculation unit 42 and the age reduction N1 up to the age of 20 that is the target of age reduction, and calculates the calculated age reduction N1. The corresponding coefficient α is read from the database A.

  The database B stores each stage (hereinafter referred to as the aging stage) N2 of the age group (aged age group) of the aging period in association with the adjustment rate γ0 corresponding to the aging stage N2. Is. A person's wrinkle change mode differs depending on the age group of the aging period or the aging period. In consideration of this, the database B of the second database 140 in the present embodiment divides the period of aging into a plurality of aging stages N2, and creates an adjustment rate γ0 for each aging stage N2. It is. In the example shown in the figure, the database B includes 45 to 40 years old, 40 to 35 years old, 35 to 30 years old, 30 to 25 years old age stages, 25 to 20 years old age stages N2 and The adjustment rate γ0 is stored correspondingly. First, the first parameter setting unit 43 first sets the adjustment rates γ0 (1), γ (2), γ (2), each stage N2 of the corresponding age group based on the registered age and the target age of 20 years as an example. Read out. Specifically, for example, when the current age is 25 years old, in the example of the database B shown in FIG. 16, the age group of the aging period corresponds only to the aging stage N2 of 25 to 20 years old. Only “1” is read as the adjustment rate γ0 (1) corresponding to the corresponding aging stage N2. On the other hand, when the current age is 44 years old, the age group of the aging period is 45 to 40 years old, 40 to 35 years old, 35 to 30 years, 35 to 30 years old, 30 to 25 years old, 25 Since it corresponds to five aging stages N2 between the ages of 20 and 20, the adjustment rate γ0 (1), the adjustment rate γ0 (2), the adjustment rate γ0 (3), and the adjustment rate corresponding to each corresponding aging stage N2. Five values of “1.1”, “1.2”, “1.15”, “1.15”, and “1” are read out as γ0 (4) and adjustment rate γ0 (5).

  The first parameter setting unit 43 calculates a final adjustment rate γ using each read adjustment rate γ0 according to the following equation (5).

γ = γ0 (1) × γ0 (2) ×... γ0 (k) (5)
γ0: Adjustment rate γ: Final adjustment rate k: Number of read adjustment rates γ0

The database C stores the aging stage N3 and the adjustment rate δ0 for each facial part in association with the aging stage N3. The human face has different wrinkle component increases with age depending on the area of the corner of the eye, forehead, jaw, etc., for example, the wrinkle component in the forehead region tends to increase greatly during the period of 30 to 40 years old, After 40 years of age, wrinkle components in the chin and the corners of the eyes tend to increase greatly. In order to reflect such a tendency, the database C provides an adjustment rate δ0 for adjusting the coefficient α for each face part in accordance with the part and the aging stage N3. The first parameter setting unit 43 confirms whether the age group of the aging period between the current age and the target age of 20 years corresponds to the aging stage N3 shown in the database C. If there is, the adjustment rate δ0 corresponding to all the corresponding aging stages N3 is read out and multiplied to obtain the final adjustment rate δ. Is determined for each part. For example, when the current age is 30 years old, the aging age group does not correspond to any aging stage N3 of the aging stage N3 of the database C shown in FIG. The On the other hand, when the current age is 45 years old, the age group of the aging period corresponds to two aging stages N3 of ~ 40 years old and 40 years old to 30 years old. Adjustment rates δ0 corresponding to the aging stage N3: “1” and “1.2” are read out, and “1.2” of 1 × 1.2 is acquired as the final adjustment rate δ of the forehead part. . For the jaw part, adjustment rates δ0: “1.2” and “1” corresponding to the respective aging stages N3 are read, and the final adjustment rate δ of the jaw part is 1.2 × 1. “1.2” is acquired. Furthermore, adjustment rates δ0 corresponding to the respective aging stages N3 are also read out for the eye corner region: “1.2” and “1”, and the final adjustment rate δ of the eye corner region is 1.2 × 1. “1.2” is acquired. For each other part, “1” is set as the final adjustment rate δ.

  Note that the second database 140 shown in FIG. 16 is an example of a woman, and in this embodiment, the second database 140 includes a database A, a database B, and a database C for male and female genders. The first parameter setting unit 43 reads each parameter from the corresponding database according to the gender of the person to be authenticated.

  In this way, the first parameter setting unit 43 reads from the database C the adjustment rate γ obtained by multiplying the coefficient α read from the database A of the second database 140 and the adjustment rate γ0 read from the database B. The adjustment rate δ for each part obtained by multiplying the adjustment rate δ0 obtained in this way is acquired, and the parameter W1 is obtained by multiplying the adjustment rate γ and the adjustment rate δ by the coefficient α to obtain the first creation execution unit 45. Output.

The first creation execution unit 45 multiplies each wrinkle component Qm extracted from the luminance component of the face image D0 by the adjustment coefficient ρ1 by the wrinkle component extraction unit 30, and multiplies the wrinkle component Qm by the adjustment coefficient ρ1. The component (adjustment component) obtained in this way is subtracted from the original image S0 (the luminance component of the face image D0) to obtain an aged image (that is, an image in which the wrinkle component is suppressed) S′1. The following formulas (6) and (7) indicate processing performed by the first parameter setting unit 43 and the first creation executing unit 45 of the authentication image creating unit 40.

ρ1 = β (S0) × W1 (7)
W1 = α × γ × δ
Where ρ1: adjustment factor
β: Pixel value dependent coefficient
W1: Parameters obtained by the first parameter setting unit 43
α: Coefficient read from database A
γ: Adjustment rate obtained by multiplying the adjustment rate γ0 read from the database B
δ: Adjustment rate obtained by multiplying the adjustment rate δ0 read from the database C

As described above, the coefficient α and the adjustment rate γ are the same for the entire original image S0, and the adjustment rate δ is set for each part of the face represented by the original image S0. It is different for each part of the face.

  The pixel value dependent coefficient β is β (S0), and is determined according to the luminance value Y0 of the pixel of the original image S0 (here, the luminance component of the face image D0). Specifically, the larger the luminance value Y0 is, the larger the value of the coefficient β used when obtaining the pixel value Y1 of this pixel. The wrinkle component Qm extracted by the wrinkle component extraction means 30 may contain components such as hair, and when creating an aged image, the hair component is about the same as the true wrinkle component. It is desirable to avoid suppression (ie subtraction). In the present embodiment, focusing on the fact that the skin portion where wrinkles or the like are generally generated is bright (that is, the luminance value is large) and the hair portion is dark (that is, the luminance value is small), the pixel having the larger luminance value has a value. True wrinkles, spots or noise can be obtained by increasing the degree of subtraction of the skin part by using the coefficient β that increases (conversely, the value becomes smaller as the pixel has a lower luminance value), weakens the degree of subtraction of the hair part, and increases the degree of subtraction of the skin part. It can be suppressed by subtracting the components such as, and the suppression of the component representing the hair is weakened.

  Next, in the authentication image creating means 40, a process for creating the second authentication image S′2 corresponding to the registered face image Dg, that is, a process by the second parameter setting unit 44 and the second creation executing unit 46. Will be described.

  The second parameter setting unit 44 uses the registered age output from the target specifying unit 3 and various parameters stored in the second database 140 for the second authentication corresponding to the registered face image Dg. The parameter W2 for creating the image S′2 (Y1) is set. Specifically, in the same manner as the first parameter setting unit 43, from the second database 140 shown in FIG. 16, from the registered age to 20 years old, the age is decreased from the registered age to the target age of 20 years. Based on the aging age N1, the aging stage N2 corresponding to the age group of the age period, and the corresponding aging stage N3, the coefficient α, the adjustment rate γ, the adjustment rate δ are obtained, and the adjustment rate γ and the adjustment rate δ are obtained. Is multiplied by the coefficient α to obtain the parameter W2 and output it to the second creation execution unit 46.

The second creation execution unit 46 multiplies the wrinkle component Qm extracted from the luminance component of the registered face image Dg by the wrinkle component extraction unit 30 by the adjustment coefficient ρ2, and multiplies the wrinkle component Qm by the adjustment coefficient ρ2. The component (adjustment component) obtained in this manner is subtracted from the original image S0 (the luminance component of the registered face image Dg) to obtain an aged image (that is, an image in which the wrinkle component is suppressed) S′2. Equations (8) and (9) below show processing performed by the second parameter setting unit 44 and the second creation executing unit 46 of the authentication image creating unit 40.

ρ2 = β (S0) × W2 (9)
W2 = α × γ × δ
Where ρ2: adjustment factor
β: Pixel value dependent coefficient
W2: parameter obtained by the second parameter setting unit 44
α: Coefficient read from database A
γ: Adjustment rate obtained by multiplying the adjustment rate γ0 read from the database B
δ: Adjustment rate obtained by multiplying the adjustment rate δ0 read from the database C

The synthesizing unit 50 determines the color difference between the pixel value Y1 of the first authentication image S′1 obtained from the luminance component of the face image D0 by the authentication image creating unit 40 and the face image D0 obtained by the YCC conversion unit 5. The first authentication image D1 corresponding to the face image D0 is obtained by combining the values Cb0 and Cr0, and the second authentication image obtained from the luminance component of the registered face image Dg by the authentication image creating means 40. The second authentication image D2 corresponding to the registered face image Dg is obtained by combining the pixel value Y1 of S′2 and the color difference values Cbg and Crg of the registered face image Dg obtained by the YCC conversion means 5, Two authentication images are output to the collation unit 70.

  The collation unit 70 of the authentication unit 100 applies the first authentication image D1 and the second authentication image D2 output from the image generation unit 60 (specifically, from the combining unit 50 of the image generation unit 60). On the other hand, pattern matching processing is performed and collation is performed, and the processing ends when the collation result is output.

  FIG. 17 is a flowchart showing processing performed in the authentication apparatus of the embodiment shown in FIG. As illustrated, when the face image D0 of the person to be authenticated obtained by the imaging unit 1 is acquired by the authentication device of the present embodiment (S10), the authentication unit 100 reads the authentication target from the first database 120. The registered face image Dg corresponding to the person, the age at the time of registration of the person corresponding to the registered face image Dg, the registration date, and the gender of the person are read (S12). And the authentication part 100 suppresses a wrinkle component with respect to the face image D0, obtains the 1st image for authentication D1, and suppresses a wrinkle component with respect to the registration face image Dg, and 2nd image for authentication D2. Is obtained (S14). The authentication unit 100 performs authentication by collating the two authentication images (S36).

 The preferred embodiment of the present invention has been described above. However, the authentication method and apparatus of the present invention and the program therefor are not limited to the above-described embodiment, and various increases and decreases may be made without departing from the gist of the present invention. Changes can be made.

  For example, in the authentication device of the embodiment described above, when suppressing the wrinkle component, the intensity of the suppression is determined according to the length of the aging period, the age group of the period, the part of the face, etc. Although it is trying to obtain a proper authentication image and improve the accuracy of authentication, it does not take into account the age of aging in particular, and it simply suppresses wrinkle components for both the registered face image and the face image at the time of authentication. Then, collation may be performed.

  Further, the method for extracting the wrinkle component and the method for suppressing the wrinkle component are not limited to the method used in the authentication apparatus of the present embodiment, and any wrinkle component can be extracted and the wrinkle component can be suppressed. A method may be used.

  Also, the method of creating the band limited image is not limited to the method of the band limited image creating unit 20 in the above-described embodiment.

  Further, in the authentication device of the above-described embodiment, paying attention to the fact that the wrinkle component is mainly present in the luminance component, the processing for suppressing the wrinkle component is performed only for the luminance component of the face image and the registered face image at the time of authentication. However, in addition to the luminance component, it may be performed for the color difference component. Since the wrinkle component suppression process performed on the color difference component can also suppress or remove the noise component in the color difference component, the accuracy of authentication can be further improved.

The block diagram which shows the structure of the authentication apparatus used as embodiment of this invention The block diagram which shows the structure of the authentication part 100 in the authentication apparatus of embodiment shown in FIG. The block diagram which shows the structure of the image generation part 60 in the authentication part 100 shown in FIG. FIG. 3 is a block diagram showing the configuration of the blurred image creation means 10 in the image generation unit 60 shown in FIG. The figure which shows the example of the one-dimensional filter F which the filtering means 12 in the blur image creation means 10 shown in FIG. 4 uses The figure which shows the process performed in the blur image creation means 10 shown in FIG. The figure which shows the frequency characteristic of the filtering image Bk produced by the filtering means 12 in the blur image creation means 10 shown in FIG. The figure which shows the example of the two-dimensional filter used by the filtering means 12 in the blur image creation means 10 shown in FIG. The figure which shows the example of the filter F1 which the interpolation means 14 in the blur image creation means 10 shown in FIG. 4 uses for interpolation of filtering image B1 The figure which shows the example of the filter F2 which the interpolation means 14 in the blur image creation means 10 shown in FIG. 4 uses for interpolation of filtering image B2 The figure which shows the frequency characteristic of the blur image Sk created by the blur image creation means 10 shown in FIG. The figure which shows the frequency characteristic of the band limited image Tk produced by the band limited image preparation means 20 in the image generation part 60 shown in FIG. The figure which shows the example of the nonlinear function f which the wrinkle component extraction means 30 in the image generation part 60 shown in FIG. 3 uses FIG. 3 is a block diagram showing the configuration of the authentication image creation means 40 in the image generation unit 60 shown in FIG. The figure which shows the content of the 1st database 120 The figure which shows the content of the 2nd database 140 The flowchart which shows the process in the authentication apparatus shown in FIG.

Explanation of symbols

Dg registered face image D1, D2 authentication image Y luminance value Cb, Cr color difference value α coefficient γ0 adjustment rate γ final adjustment rate δ0 adjustment rate for each region δ final adjustment rate for each region β pixel value dependent coefficient ρ1 , Ρ2 Adjustment factor

Claims (9)

In an authentication method for performing authentication using an actual age face image that is a face image at the time of authentication of a person to be authenticated and a registered face image registered at a predetermined age of the person,
Each of the current age face image and the registered face image is subjected to a wrinkle component suppression process to create a corresponding authentication image,
The authentication is performed by comparing the two authentication images .
The wrinkle component suppression treatment is
Based on the target image to be the current face image or the registered face image, create a plurality of band limited images representing components for each of a plurality of frequency bands of the target image,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to obtain the pixel values of multiple converted images as wrinkle components,
Multiplying a pixel value of the plurality of converted images by a predetermined adjustment coefficient to obtain a pixel value of the adjustment component image;
An authentication method, which is a process of suppressing a wrinkle component by subtracting a pixel value of the adjustment component image from a pixel value of the target image .   The authentication method according to claim 1, wherein the wrinkle component suppression process is performed on the current face image with a suppression strength stronger than the suppression strength of the wrinkle component suppression process on the registered face image. The adjustment factor, authentication method according to claim 1 or 2, wherein said at one determined according to the pixel values of the target image. An authentication device that performs authentication using an actual-age face image that is a face image at the time of authentication of a person to be authenticated, and a registered face image registered at a predetermined age of the person,
A wrinkle component suppression unit that performs a wrinkle component suppression process on the current face image and the registered face image, respectively, and creates corresponding authentication images,
Collating means for performing the authentication by collating the two images for authentication obtained by the wrinkle component suppressing means ,
The wrinkle component suppressing means is
Based on the target image to be the current face image or the registered face image, create a plurality of band limited images representing components for each of a plurality of frequency bands of the target image,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to obtain the pixel values of multiple converted images as wrinkle components,
Multiplying a pixel value of the plurality of converted images by a predetermined adjustment coefficient to obtain a pixel value of the adjustment component image;
An authentication apparatus , wherein the pixel value of the adjustment component image is subtracted from the pixel value of the target image to suppress a wrinkle component . The wrinkle component suppression means performs the wrinkle component suppression process on the current face image with a suppression strength stronger than the suppression strength of the wrinkle component suppression processing on the registered face image. Item 5. The authentication device according to Item 4 . The authentication apparatus according to claim 4 or 5 , wherein the adjustment coefficient is determined according to a pixel value of the target image. A program for causing a computer to execute an authentication process for performing authentication using an actual face image that is a face image at the time of authentication of a person to be authenticated and a registered face image registered at a predetermined age of the person,
The authentication process performs a wrinkle component suppression process on the actual face image and the registered face image, respectively, and creates a corresponding authentication image.
Do and a process for matching two of said authentication image Ri,
The wrinkle component suppression treatment is
Based on the target image to be the current face image or the registered face image, create a plurality of band limited images representing components for each of a plurality of frequency bands of the target image,
For an input value in which the absolute value of the output value is less than or equal to the absolute value of the input value and the absolute value is less than or equal to a predetermined threshold value, the absolute value of the output value increases as the absolute value of the input value increases. On the other hand, for an input value whose absolute value is larger than the predetermined threshold value, nonlinear conversion processing in which the absolute value of the output value is equal to or smaller than the absolute value of the output value corresponding to the predetermined threshold value is performed for each band limitation. Apply to each pixel value of the image to obtain the pixel values of multiple converted images as wrinkle components,
Multiplying a pixel value of the plurality of converted images by a predetermined adjustment coefficient to obtain a pixel value of the adjustment component image;
The pixel values of the adjustment component image, the program characterized by suppressing process der Rukoto wrinkles component is subtracted from the pixel values of the target image. The program according to claim 7 , wherein the suppression strength of the wrinkle component suppression process for the current face image is stronger than the suppression strength of the wrinkle component suppression process for the registered face image. 9. The program according to claim 7 , wherein the adjustment coefficient is determined according to a pixel value of the target image.

Images