JP2011115393A - Skin feature determination system, skin feature determination method, and skin feature determination program - Google Patents

Abstract

It is possible to discriminate human characteristics of a subject relatively easily on the basis of generalized feature values, accurately calculate a fractal dimension, and, as a result, subject's age (skin To provide a skin feature discrimination system and a skin feature discrimination method that can accurately perform human body feature discrimination such as age), gender, and presence / absence of abnormal skin condition.
A skin feature determination system S can extract a skin feature as a quantitative feature amount relatively easily by calculating a fractal dimension in an image of a skin area A of the subject, and obtains the acquired subject. A statistical process is performed using a plurality of original images in the skin region A and a divided image obtained by dividing the original image, and an operation for eliminating the influence of artifacts existing on the skin image is performed. .
[Selection] Figure 1

Description

  The present invention relates to a skin feature discrimination system and a skin feature discrimination method for discriminating human characteristics of a subject such as age or sex using a skin image.

  In recent years, various methods or systems have been proposed for quantifying a human characteristic of a subject from the viewpoint of security, and for identifying and identifying the human characteristic of the subject. For example, a fingerprint, a voiceprint, a cornea, A personal authentication system based on the acquired unique information has been put into practical use by acquiring unique information about the human body such as a vein (hereinafter simply referred to as “unique information”). Such a personal authentication system exhibits a high security capability capable of distinguishing an individual person from other persons, while having to register the individual person's unique information in advance, and the unique information is a simple device. Or, it is difficult to register by a method, and strict management is required for the registered information.

  On the other hand, a skin diagnosis system using a surface image of the skin of a human body has been proposed while interest in beauty has increased in recent years (see, for example, Non-Patent Document 1). However, regarding skin diagnosis, since there is a high correlation with skin age or skin condition, feature quantities that can be measured relatively easily have not been established.

  In such a situation, the inventors can use it in the security field or the medical and beauty field, and it is relatively easy based on the generalized feature amount without registering individual specific information in advance. Has invented an apparatus (hereinafter referred to as “skin feature discrimination system”) that can perform human body feature discrimination of a subject.

  This skin feature discrimination system calculates a fractal dimension based on a plurality of skin images of a subject, compares the calculated fractal dimension with a standard fractal dimension, and based on the comparison result, the age of the subject (skin age) Or gender is estimated (patent document 1).

Yoshitake Takemae, Hideo Saito, Shinji Ozawa: “Skin Diagnosis System Using Skin Surface Images”, Transactions of the Society of Instrument and Control Engineers, Vol. 37, no. 11, pp. 1097-1103 (2001)

Japanese Patent Laid-Open No. 2007-50158

  However, in the skin feature discrimination system described in Patent Document 1, the fractal dimension based on the skin image is calculated using the box counting method, and the age (skin age) of the subject cannot be accurately estimated. There is a case. That is, in this skin feature determination system, in the skin image that is the basis of the calculation, if there are many artifacts present on the skin image due to body hair, dust, etc. of the imaging part, the fractal dimension is accurately calculated. In some cases, the estimated age (skin age) or gender of the subject deviates from the actual one.

  Therefore, the present invention has been made to solve the above problems, and its purpose is to perform human feature determination of a subject relatively easily based on a generalized feature amount called a fractal dimension. It is possible to accurately calculate the fractal dimension, and as a result, accurately determine the human characteristics of the subject such as the subject's age (skin age), gender, and the presence or absence of abnormal skin conditions. An object is to provide a skin feature discrimination system, a skin feature discrimination method, and a skin feature discrimination program.

  (1) In order to solve the above problems, the present invention provides an acquisition unit that acquires a plurality of images of a predetermined region on the skin of a subject as skin image data, and an image of each original image formed by the acquired skin image data Dividing means for dividing the region into a plurality of parts and generating divided images formed from the plurality of partial images; fractal dimension calculating means for calculating the original images and the fractal dimensions of the partial images; First representative value calculation for changing the number of image data and calculating a representative value of the fractal dimension in the original image as a first representative value based on the calculated fractal dimension in each original image. And the number of data of the divided images used for the calculation are changed, and for each number of the data, the calculated fractal dimension in each partial image for each of the divided numbers. Therefore, a second representative value calculating means for calculating a representative value of the fractal dimension in the divided image as a second representative value, each calculated first representative value, and each calculated second representative for each of the divided numbers. A function calculation means for calculating a first function with the number of data and the first representative value as variables and a second function for each of the number of divisions with the number of data and the second representative value as variables, based on values; Specific means for specifying one fractal dimension determined based on the calculated first function and each second function as a specific value, and a predetermined fractal dimension indicating a standard human body characteristic as a reference value are stored in advance. A database, a comparison unit that compares the specified value and the reference value, and a determination unit that determines a human characteristic of the subject who acquired the skin image data based on the comparison result. Preparation It has a configuration.

  Usually, the skin of a subject has self-similarity for a part of a predetermined region and the entire region because of its characteristics. Also, by using the concept of fractal dimension, that is, by calculating the fractal dimension in the shape of a predetermined region, even a shape (figure) that does not have a characteristic length has self-similarity. In the case of a shape (figure), it is possible to analyze the feature quantity quantitatively. On the other hand, when image analysis is performed on a predetermined region of the subject's skin using the fractal dimension, it is difficult to accurately calculate the fractal dimension due to artifacts caused by body hair, sebum, attached dust, or the like.

  Therefore, according to the above-described configuration, the present invention extracts a feature of the subject's skin as a quantitative feature amount by calculating a fractal dimension in an image of a predetermined region of the subject's skin, and performs feature determination. The fractal dimension of the skin image can be calculated while statistically eliminating the influence of artifacts existing on the skin image by using the plurality of original images on the acquired subject's skin and the divided images obtained by dividing the original image. Therefore, it is possible to accurately determine the human characteristics of the subject based on the skin image such as the age (skin age) of the subject, sex, and the presence / absence of an abnormal skin condition.

  (2) Moreover, in this invention, the said discrimination means has the structure which discriminate | determines the human body characteristic of the test subject who acquired the said skin image data of at least any one of age or sex based on the said comparison result. Is preferred.

  With this configuration, the present invention can accurately calculate the fractal dimension of the skin image while eliminating the influence of artifacts statistically generated on the image using a plurality of original images and divided images obtained by dividing the plurality of original images. Therefore, the age (skin age) and sex of the subject can be accurately determined.

  (3) Moreover, in this invention, the said discrimination means has the structure which discriminate | determines the presence or absence of the said normal state of the skin as a human body characteristic of the test subject who acquired the said skin image data based on the said comparison result. Is preferred.

  With this configuration, the present invention can accurately calculate the fractal dimension of the skin image while eliminating the influence of artifacts statistically generated on the image using a plurality of original images and divided images obtained by dividing the plurality of original images. Therefore, the presence or absence of an abnormal skin condition can be accurately determined.

  (4) Further, in the present invention, the dividing unit has a configuration for generating a plurality of divided images obtained by dividing each of the original images into a plurality of different numbers by changing the size of the divided areas used for the division. Is preferred.

  When artifacts are present on the original image, if the number of divisions is changed, partial images that are not affected by artifacts appear as the number of divisions increases. Naturally, the fractal dimension of the image without the influence of the artifact accurately indicates the feature amount of the skin of the subject.

  Therefore, according to the present invention, by using the plurality of original images and the divided images obtained by dividing the plurality of original images, the present invention can eliminate artifacts and calculate an accurate fractal dimension in the skin of the subject.

  (5) Moreover, in this invention, it is preferable that the said acquisition means has the structure which acquires the several skin image of a different skin area | region as skin image data.

  With this configuration, the present invention can calculate the fractal dimension in various skin regions, thereby eliminating the influence of unique artifacts in each skin region and accurately calculating the fractal dimension in the subject's skin. can do.

  (6) Moreover, in this invention, it is preferable that the said acquisition means has the structure which acquires the several skin image of the same skin area | region as skin image data.

  With this configuration, the present invention can eliminate the influence of temporal artifacts existing in the skin region or the artifacts generated at the time of imaging, so that the calculated fractal dimension can be calculated. The fractal dimension in the skin can be calculated accurately.

  (7) Further, in the present invention, the fractal dimension calculation means calculates the fractal dimensions of the original image and the partial images based on the shapes and gray scales in the original image area and the divided image area, respectively. It is preferable to have a configuration to calculate.

  With this configuration, the present invention can calculate the fractal dimension by performing a three-dimensional analysis using the shape and gray scale in the original image region and the divided image region, so that each original image and each partial image can be calculated. The calculation accuracy of the fractal dimension can be improved.

  (8) Moreover, in this invention, it is preferable that the said fractal dimension calculation means has the structure which each calculates the fractal dimension of each said original image and each said partial image by the said power spectrum method.

  With this configuration, the present invention can reduce the computation load by using a simple formula for three-dimensional analysis of the shape and gray scale in the original image region and the divided image region. The fractal dimension in the image can be calculated accurately and at high speed.

  (9) Moreover, in this invention, it further has an imaging means which images the predetermined area | region in the said test subject's skin, and the said acquisition means has the structure which acquires the image of the imaged predetermined area | region as skin image data. Is preferred.

  With this configuration, since the present invention has an imaging means, each process can be performed by an integrated device from skin image acquisition until human body characteristics are determined, so connection with other devices, etc. Without performing complicated operations, it is possible to accurately determine the human characteristics of the subject such as the subject's age (skin age), sex, and the presence or absence of an abnormal skin condition.

  (10) Moreover, in this invention, it further has a communication means connected with the said database via a public network line, and the said database is fixedly installed in the place different from the place where a skin characteristic discrimination system is used. It preferably has a configuration.

  With this configuration, the present invention can perform human characteristic determination of a subject using a database that is fixedly installed remotely from a place from which the subject's image data is acquired. In addition, if only the reference value in the database is updated, the latest data can be provided to each feature discriminating apparatus in a plurality of skin feature discriminating systems. Therefore, the present invention can be realized easily and at low cost.

  (11) In order to solve the above-described problem, the present invention provides an acquisition step of acquiring a plurality of images of a predetermined region on the subject's skin as skin image data, and an image of each original image formed by the acquired skin image data A division step of dividing a region into a plurality of parts to generate a divided image formed from a plurality of partial images, a fractal dimension calculation step of calculating each of the original image and the fractal dimension of each of the partial images, and an original used for calculation First representative value calculation for changing the number of image data and calculating a representative value of the fractal dimension in the original image as a first representative value based on the calculated fractal dimension in each original image. The fractal dimension in each of the calculated partial images for each of the division numbers is changed for each number of data by changing the number of data of the divided images used for the process and calculation. A second representative value calculating step of calculating a representative value of a fractal dimension in the divided image as a second representative value based on the calculated first representative value and the calculated second representative value for each of the divided numbers. A function calculation step of calculating a first function using the number of data and the first representative value as a variable and a second function for each of the number of divisions using the number of data and a second representative value as a variable based on a value; A specifying step of specifying one fractal dimension determined based on the calculated first function and each of the second functions as a specific value, the specified specific value, and a fractal dimension as a reference value stored in advance. It has a configuration including a comparison step for comparison, and a determination step for determining the human body characteristics of the subject who acquired the skin image data based on the comparison result.

  With this configuration, the present invention obtains the acquired subject when performing feature determination by extracting the feature of the subject's skin as a quantitative feature amount by calculating the fractal dimension in the image of the predetermined region of the subject's skin. It is possible to calculate the fractal dimension of the skin image while statistically eliminating the influence of artifacts existing on the skin image by using a plurality of original images on the skin and the divided images obtained by dividing the original image. It is possible to accurately determine the human characteristics of the subject based on the skin image such as the age (skin age), sex, and presence / absence of abnormal skin condition.

  (12) In order to solve the above-described problem, the present invention determines a human body characteristic of a subject based on skin image data from a plurality of images obtained by a computer imaging a predetermined region on the subject's skin. A dividing program that divides the image area of each original image formed by the acquired skin image data into a plurality of parts, and generates a divided image formed from a plurality of partial images; Fractal dimension calculating means for calculating the fractal dimension of each original image and each partial image, and changing the number of data of the original image used for the calculation, and for each data number based on the calculated fractal dimension in each original image A first representative value calculating means for calculating a representative value of a fractal dimension in the original image as a first representative value; A second representative that changes the number of data and calculates a representative value of the fractal dimension in the divided image as a second representative value based on the calculated fractal dimension in each partial image for each of the divided numbers. Based on the calculated first representative value and the calculated second representative value for each division number, the first function and the data using the number of data and the first representative value as variables A function calculation means for calculating each second function for each of the division numbers using the number and the second representative value as variables, and a specific fractal dimension determined based on the calculated first function and each second function as a specific value A specifying means for specifying, a comparing means for comparing the specified specific value and a fractal dimension as a reference value stored in advance, a person of the subject who has acquired the skin image data based on the comparison result Have characteristics determining means for determining a configuration to function as a.

  With this configuration, the present invention obtains the acquired subject when performing feature determination by extracting the feature of the subject's skin as a quantitative feature amount by calculating the fractal dimension in the image of the predetermined region of the subject's skin. It is possible to calculate the fractal dimension of the skin image while statistically eliminating the influence of artifacts existing on the skin image by using a plurality of original images on the skin and the divided images obtained by dividing the original image. It is possible to accurately determine the human characteristics of the subject based on the skin image such as the age (skin age), sex, and presence / absence of abnormal skin condition.

  The present invention calculates the fractal dimension in an image of a predetermined region of the subject's skin to extract the feature of the subject's skin as a quantitative feature quantity, and performs feature discrimination, thereby obtaining a plurality of acquired skins of the subject's skin. The fractal dimension of the skin image can be calculated while statistically eliminating the influence of artifacts existing on the skin image using the original image and the divided image obtained by dividing the original image. It is possible to accurately determine the human characteristics of the subject based on the skin image, such as age), sex, and the presence / absence of an abnormal skin condition.

It is a block diagram which shows the structure in 1st Embodiment of the skin diagnostic apparatus which concerns on this invention. It is an example of the skin image (original image) used when discriminating the human body characteristic of a test subject. It is an example of the list | wrist of the standard fractal dimension as a reference value memorize | stored in the database of 1st Embodiment. It is an example of the divided image in 1st Embodiment. It is an example which shows the list of the calculated fractal dimension of each partial image in the 8-part dividing image in 1st Embodiment. It is a figure for demonstrating calculation of the representative value in 1st Embodiment. It is a figure for demonstrating calculation of the specific value in 1st Embodiment. It is a flowchart which shows the determination process which determines only the test subject's age and sex, or sex including the calculation process which calculates the fractal dimension of 1st Embodiment. It is a flowchart which shows the determination process which determines the presence or absence of skin abnormality including the calculation process which calculates the fractal dimension in 1st Embodiment. It is a block diagram which shows the structure in 2nd Embodiment of the skin diagnostic apparatus which concerns on this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the embodiment described below, the skin feature determination system, the skin feature determination method, and the skin feature determination program of the present invention are applied to a skin diagnostic apparatus that determines the human body characteristics of the subject based on the skin image of the subject. It is applied.

[First Embodiment]
First, a first embodiment of a skin feature determination system S according to the present invention will be described with reference to FIGS.

(System configuration)
First, the configuration of the skin feature determination system S of the present embodiment will be described with reference to FIGS. FIG. 1 is a block diagram showing the configuration of the skin feature discrimination system S in the present embodiment, and FIG. 2 is an example of a skin image (original image) used when discriminating human characteristics of the subject. . FIG. 3 is an example of a standard fractal dimension as a reference value stored in the database 200 of this embodiment.

  As shown in FIG. 1, the skin feature determination system S of the present embodiment includes an imaging camera device 100 that captures a predetermined region (hereinafter also referred to as “skin region”) A of the subject's skin, and the human body of the subject. A database 200 storing data (fractal dimension as a reference value to be described later) used for distinguishing features, and image data of the skin region A imaged by the imaging camera device 100 (hereinafter referred to as “skin image data”). .), And a discrimination device 300 that discriminates the human characteristics of the subject such as age (skin age), sex, and presence / absence of abnormal skin based on the calculated fractal dimension, and discrimination The display device 400 displays the determination result determined by the device 300, and the operation unit 500 that performs other operations such as the imaging camera device 100.

  This skin feature discriminating system S is an image formed by each skin image data with a fractal dimension indicating a numerical index that can quantitatively represent the “complexity” of a figure, phenomenon or state having no characteristic length. And the subject's age (skin age), sex, and presence / absence of abnormal skin condition by comparing the calculated fractal dimension with a standard value, that is, a standard fractal dimension serving as a reference value. It is designed to discriminate human characteristics.

  Usually, human skin such as a subject has a characteristic of self-similarity in which a shape formed in a part of a predetermined region and a shape formed by the entire region are similar. Also, by using the concept of fractal dimension, that is, even if the figure does not have a characteristic length by calculating the fractal dimension, It is possible to analyze the feature amount. On the other hand, when image analysis is performed on the subject's skin using the fractal dimension, it is difficult to accurately calculate the fractal dimension due to artifacts due to body hair, sebum, attached dust, or the like. That is, for example, as shown in FIG. 2 (a), a skin image without an artifact is ideal, but such a skin image is rarely obtained, and usually as shown in FIG. 2 (b). Since the self-similarity in the image is disturbed by the body hair H or the like, it is difficult to accurately calculate the fractal dimension.

  Therefore, the skin feature determination system S according to the present embodiment can extract the skin feature as a quantitative feature amount relatively easily by calculating the fractal dimension in the image of the skin region A of the subject, and is acquired. The calculation is performed to eliminate the influence of the artifacts existing on the skin image by performing statistical processing using the plurality of original images in the skin area A of the subject and the divided images obtained by dividing the original image. ing. Therefore, since the skin feature determination system S of the present embodiment can accurately specify the fractal dimension of the skin image, the subject based on the skin image, such as the subject's age (skin age), gender, and the presence or absence of an abnormal skin condition It is possible to accurately perform human body feature discrimination.

  Specifically, the imaging camera device 100 of the present embodiment has a predetermined imaging element such as a CCD (Charge Coupled Device Image Sensor), is movably or fixedly installed, and is skinned at a predetermined part of a subject. This is used to image a predetermined area in a plurality of times. The imaging camera device 100 is driven and controlled by the determination device 300 such as zoom, focus, adjustment of the imaging position, and the number of times of imaging. The captured skin image is converted into skin image data of a predetermined format and output to the determination device 300. It is supposed to be. In particular, the imaging camera device 100 captures the same skin region A a plurality of times or images different skin regions A at different sites a plurality of times, and outputs the captured skin images to the determination device 300 as skin image data. It is like that.

  In the database 200 of this embodiment, a predetermined fractal dimension indicating the characteristics of a standard human body P as a reference value is stored in advance. Specifically, this database 200 includes a standard fractal dimension for each age and gender that is used when the age of the subject is a determination target, a male and a male regardless of the age used when the subject's gender is a determination target. A standard fractal dimension of a woman and a standard fractal dimension used when the presence or absence of an abnormal state of the subject's skin is determined are stored. For example, as shown in FIG. 3, the database 200 stores standard fractal dimensions as reference values for men and women in their 20s to 40s.

  In the reference value stored in the database 200, as shown in FIG. 3, a predetermined range is also stored. If the specific value X belongs to the range, the age indicating the range is stored. And it is determined to be gender. In addition, as a fractal dimension as a reference value used when determining only the sex of male or female, for example, the fractal dimension of male “1.3408 ± 0.11” and female “1.4758 ± 0.18” is used. It is stored in the database 200. In the present embodiment, as the fractal dimension as the reference value stored in the database 200, a value obtained in advance by an experiment or other method is used.

  The discrimination device 300 according to the present embodiment is mainly configured by a central processing unit (CPU), and comprehensively controls the overall functions of the skin feature discrimination system S. In particular, the determination device 300 controls the imaging camera device 100 and receives skin image data output from the imaging camera device 100. The discriminating apparatus 300 uses a fractal dimension of each image (hereinafter referred to as “original image”) formed by a plurality of input skin image data and a plurality of partial images obtained by dividing the original image into predetermined regions. The fractal dimension is calculated for each of the divided images. Further, the discriminating apparatus 300 compares the calculated fractal dimension with the fractal dimension as a reference value indicating the human characteristics of the subject to be discriminated stored in advance in the database 200, and based on the comparison result. Thus, the human characteristics of the subject are discriminated.

  Specifically, as shown in FIG. 1, the determination apparatus 300 performs an input interface 310 that performs predetermined processing on captured skin image data, and divides an original image formed by each skin image data. A division processing unit 320 to perform, a fractal dimension calculation unit 330 for calculating a fractal dimension in a plurality of original images and an image obtained by dividing the original image, and specifying a fractal dimension in the subject, and a fractal dimension as a standard reference value And a comparison / determination unit 340 that determines the human body characteristics of the subject.

  Further, the determination device 300 includes a camera control unit 350 that controls the imaging camera device 100 based on an operation of the operation unit 500, a display control unit 360 that controls the display device 400, and an arithmetic process that calculates a fractal dimension in the subject. The overall control unit 370 controls each unit by a determination process for determining a human characteristic of the subject and other control processes, and a ROM / RAM 380 in which predetermined data is stored. The details of the operation of each part of the determination apparatus 300 according to the present embodiment will be described later.

  The display device 400 of the present embodiment is configured by a CRT (Cathode Ray Tube) liquid crystal element or an EL (Electro Luminescence) element. Specifically, the display device 400 displays a predetermined image based on display data generated by the display control unit 360 under the control of the determination device 300. For example, the display device 400 includes a display related to the operation of the imaging camera device 100, a captured skin image, captured skin image data, a subject's age (skin age), gender, and presence / absence of an abnormal skin condition. A predetermined image necessary for the determination device 300 to perform the above operation, such as a determination result of a characteristic feature, is displayed.

  The operation unit 500 according to the present embodiment includes a large number of keys such as operation buttons for inputting each operation command, ten keys, and other numeric keys. The operation unit 500 performs determination processing type determination, operation of the imaging camera device 100 via the camera control unit 350, operation for displaying human characteristics on the display device 400, and other input operations. It has come to be used.

(Determination device)
Next, details of each part of the determination device 300 of this embodiment will be described with reference to FIGS. 1 and 4 described above. FIG. 4 is an example of a divided image in the present embodiment.

  As described above, the determination device 300 includes the input interface 310, the division processing unit 320, the fractal dimension calculation unit 330, the comparison determination unit 340, the display control unit 360, and the camera control unit as illustrated in FIG. 350, a ROM / RAM 380, and an overall control unit 370.

  The input interface 310 receives a plurality of skin image data output from the imaging camera device 100, and executes predetermined processing on each received skin image data. For example, the input interface 310 calculates data name assignment, imaging date and time, imaging order, data format conversion, image enlargement, reduction, rotation, and other fractal dimensions for the received skin image data. Therefore, processing necessary for this is selectively and automatically executed under the control of the overall control unit 370.

  The division processing unit 320 equally divides an image area of a skin image (original image) formed by each skin image data subjected to predetermined processing by the input interface 310 into n image areas (hereinafter referred to as “partial image areas”). )). Further, the division processing unit 320 changes the number of divisions of an image formed from a plurality of partial image regions (that is, partial images) (hereinafter referred to as “divided images”) for each original image. (For example, the number of divisions is changed from 2 to 8). That is, the division processing unit 320 equally divides the image area of each original image into two, three, four,. . . . Each of the divided images divided into n (n is an integer) is generated. For example, the division processing unit 320 according to the present embodiment generates each divided image from two divisions to eight divisions. Specifically, when generating an eight division image, the division processing unit 320 The one image shown in FIG. 2B is equally divided into eight parts for each predetermined area as shown in FIG.

The fractal dimension calculation unit 330 calculates, for example, the fractal dimension of each original image and the fractal dimension of each partial image constituting each divided image using the power spectrum method. Then, the fractal dimension calculation unit 330 changes the number of images (that is, the number of data) by a predetermined calculation process so as to change the number of data and the number of divisions (with the original image as the division number 1) The fractal dimension in the divided image is calculated, and one fractal dimension determined based on the calculated data number and the fractal dimension (hereinafter referred to as “representative value”) V for each divided number is specified as the specific value X. It has become.
Details of the calculation of the fractal dimension of the fractal dimension calculation unit 330 in this embodiment will be described later.

  When the comparison determination unit 340 executes determination processing of either the age determination of the subject (hereinafter simply referred to as “age determination”) or the determination of the sex of the subject (hereinafter simply referred to as “sex determination”). In addition, the corresponding fractal dimension (reference value) is read from the database 200. In addition, the comparison determination unit 340 compares the read plural or one fractal dimension (reference value) with the specific value X of the fractal dimension calculated by the calculation unit, and the reference value read by the specific value X Are included in a predetermined range centered on, only the characteristics of the reference value, that is, the age and gender or sex to which the reference value belongs are determined as the determination results. The comparison determination unit 340 outputs the determined determination result to the display device 400 via the display control unit 360.

  On the other hand, when the comparison determination unit 340 determines whether or not the subject has an abnormal skin condition (hereinafter simply referred to as “skin abnormality determination”), the age of the subject input in advance via the operation unit 500 is used. And the fractal dimension that matches the gender is read from the database 200 and compared with the specific value X. When the specific value X is less than 50% or 150% or more of the reference value, it is determined that the skin is abnormal, and the specific value Is 50% or more of the reference value and 150% or less, it is determined that there is no abnormality in the skin. The comparison determination unit 340 also outputs the determined determination result to the display device 400 via the display control unit 360 for the skin abnormality determination.

  In addition, the comparison determination unit 340 of the present embodiment is not limited to the above determination method. For example, the specific value X is within a range stored in advance in the database 200 around the reference value (for example, if it is a male in his 20s, 13408 ± 0.11) or out of range, if it is outside this range, it is determined that there is an abnormality in the skin, and if within this range, it is determined that there is no abnormality in the skin You may do it.

  The camera control unit 350 is configured to control the imaging camera device 100 in accordance with an input operation of the operation unit 500. For example, as described above, the camera control unit 350 performs zoom adjustment, focus adjustment, imaging position adjustment, and the like when the imaging camera device 100 images the skin of a predetermined part of the subject according to the input operation of the operation unit 500. Various operations such as setting of the number of times of imaging are controlled.

  Under the overall control unit 370, the display control unit 360 displays a display relating to the operation of the imaging camera device 100, a captured skin image, a predetermined image related to the captured skin image data, a discrimination result of human characteristics, and the like Display data for displaying on the display device 400 a predetermined image necessary for the operation of the above 300 is generated.

  The overall control unit 370 is connected to each unit by the bus B, and controls each unit by the determination process for determining the human body characteristics of the subject and the various control processes including the calculation process for calculating each fractal dimension. The skin feature discrimination system S is functionally operated.

  The ROM / RAM 380 stores various programs for executing the above arithmetic processing, programs for executing the above determination processing, and various control programs, and temporarily stores predetermined data during the operation of the various programs. Retained.

(Fractal dimension calculator)
Next, the details of the fractal dimension calculation unit 330 of the present embodiment will be described with reference to FIGS. FIG. 5 is an example showing a list of calculated fractal dimensions of each partial image in the eight-divided image in the first embodiment, and FIG. 6 is for explaining the calculation of the representative value V in the present embodiment. FIG. FIG. 7 is a diagram for explaining the calculation of the specific value X in the present embodiment.

  As described above, the fractal dimension calculation unit 330 of the present embodiment uses the power spectrum method as a fractal dimension calculation method. Normally, when the power spectrum method is applied to an image, when a two-dimensional Fourier transform is performed on an image of a target two-dimensional plane and a power spectrum S (f) is obtained, the spectrum S (f) is Fourier-transformed. Is the square of the absolute value of the result obtained. When the spectrum S (f) is in the form of a power, the exponent (β) and the fractal dimension (D) of the power satisfy the relationship of (Equation 1).

  Therefore, the fractal dimension calculation unit 330 of this embodiment calculates the power spectrum S (f) for each partial image in each original image and each divided image, and (Equation 1) is calculated for each power spectrum S (f). It is used to calculate the fractal dimension in each partial image in each original image and each divided image.

  Note that the fractal dimension calculated in the present embodiment is a dimension of one dimension or more and two dimensions or less. As for the power spectrum method, for example, A. P. Araghy, M.M. Sato and M.C. Kasuga's "EFEICENT FRANCTAL DIMENSION DIMENSION CALCULATION METHOD FOR FEATUE EXTRCTION OF STIN IMAGES"

  On the other hand, when the fractal dimension of each original image is calculated, the fractal dimension calculation unit 330 changes the number of data and calculates one fractal dimension (representative value V) for each number of data. For example, the fractal dimension calculation unit 330 is referred to as an original image (hereinafter referred to as “first original image”) formed from the first captured skin image data in the order of imaging provided by the input interface 310 unit. The fractal dimension calculated for the original image formed by the nth skin image data is referred to as “nth original image”) is set as the representative value V of the fractal dimension when the number of data is “1” ( Calculation). In addition, the fractal dimension calculation unit 330 calculates the average of the fractal dimensions calculated for the first original image and the second original image as the representative value V when the number of data is “2”. The fractal dimension calculation unit 330 then calculates the fractal dimension of each fractal dimension calculated from the first original image to the nth original image when the number of data is “3” or later, as in the case of the number of data “2”. An average is calculated, and the calculated average is calculated as a representative value V of the fractal dimension in the number of data. Note that the fractal dimension calculation unit 330 according to the present embodiment calculates the representative value V of the fractal dimension up to the number of data “8”, for example.

  On the other hand, when the fractal dimension of each partial image in each divided image is calculated, the fractal dimension calculation unit 330 calculates one fractal dimension for each number of divisions and for each number of data. Specifically, when the number of data in the calculated fractal dimension is “n” and the number of divisions is “m”, the fractal dimension calculation unit 330 (a) m-division image (hereinafter, “ Hereinafter, the m-divided image of the n-th original image is referred to as “the n-th divided image”) (ie, the first partial image or the second partial image). (B) each fractal dimension in each partial image calculated for the second m-division image, and (c) each fractal dimension in each partial image calculated for the third m-division image,. . . . And (d) comparing each fractal dimension in each partial image calculated for the nm-th divided image and approximating the fractal dimension in the other divided images in each divided image (for example, for the first two-divided image) Fractal dimension of the first partial image, fractal dimension of the second partial image and the like for the second divided image) are extracted for each of the first to n-th divided images, and each fractal dimension of each of the extracted divided images is extracted. Is set as the representative value V of the fractal dimension of the m-division image for the number of data n.

For example, when the number of data is “6” and the number of divisions is “8”, the fractal dimension calculation unit 330 sets the fractal dimension of each partial image in each of the eight divided images as shown in FIGS. Based on this, a fractal dimension that can be approximated with other divided images is extracted in each divided image, that is,
(A) a sixth partial image in the first eight-divided image;
(B) a first partial image in the second eight-divided image;
(C) a fifth partial image in the third eight-divided image;
(D) a first partial image in the fourth eight-divided image;
(E) Fractal dimensions of the sixth partial image in the fifth eight-divided image, and (f) Fractal dimension of the fifth partial image in the sixth eight-divided image, and the average of the fractal dimensions in the extracted eight-divided images is calculated. It is calculated as a representative value V (fractal dimension) when the number of data is “6” and the number of divisions is “8”.

  When the number of data is “1”, for example, the average of each fractal dimension in the 1m divided image is set to the representative value V. In the calculation of the representative value V, regarding the extraction of the fractal dimension that can be approximated for each divided image, it is only necessary to extract each fractal dimension that can be approximated to each of the divided images for each divided image. Calculated by arithmetic processing.

  In addition, when the representative value V calculated for each number of data and for each number of divisions is calculated, the fractal dimension calculation unit 330 of the present embodiment calculates the number of data and the representative value V based on these representative values V. A function as a variable is calculated, and one fractal dimension determined based on each calculated function is specified as a final specific value X as a fractal dimension. Specifically, as shown in FIG. 7, the fractal dimension calculation unit 330 calculates the function of the fractal dimension at the representative value V calculated for each number of data for each number of divisions from “1 to 8”. Each is calculated, and the fractal dimension of the intersection where the functions of the respective division numbers intersect on the graph is specified as the specific value X.

  As described above, the calculation unit of the present embodiment statistically eliminates the influence of artifacts existing on the skin image, and accurately identifies the fractal dimension of the skin image in order to obtain a plurality of originals on the acquired subject's skin. One fractal dimension in the skin image data of the subject is specified using the divided image obtained by dividing the image and the original image.

(Judgment processing (I))
Next, in the present embodiment, a determination process for determining only the age and sex or gender of a subject including a calculation process for calculating a fractal dimension based on skin image data acquired by imaging will be described with reference to FIG. In addition, FIG. 8 is a flowchart which shows the determination process which determines only a test subject's age and sex, or sex including the calculation process which calculates the fractal dimension in this embodiment.

  In this operation, the standard fractal dimension of each age and gender used when the subject's age is already determined in the database 200 is stored in the database 200, and regardless of the age used when the subject's gender is determined. It is assumed that standard fractal dimensions are stored in advance. In addition, it is assumed that a predetermined region of the skin of the subject has already been captured by the imaging camera device 100 and skin image data has been acquired.

  First, when the user's operation of the operation unit 500 is set to determine the subject's age and gender or only gender as a human characteristic of the subject (step S101), the overall control unit 370 divides into the division processing unit 320. By changing the number, the original image formed by the skin image data is evenly generated into a plurality of divided images from 2 to n (step S102). For example, as described above, the division processing unit 320 generates respective divided images from 2 divisions to 8 divisions.

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate the fractal dimension of each partial image in the original image and each divided image, and calculates the representative value V of the fractal dimension for each number of data and for each number of divisions ( Step S103).

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate a fractal dimension function with the number of data as a variable for each division number based on the representative value V of the fractal dimension calculated for each number of data and for each division number. Generate (step S104).

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate the intersection point on the graph in each function as the specific value X based on the fractal dimension function of each division number (step S105). At this time, when the functions of the fractal dimension do not intersect at one point, the function having the maximum number of functions at the intersection is calculated as the specific value X.

  Next, the overall control unit 370 causes the comparison determination unit 340 to read the reference value of the fractal dimension for each age and gender stored from the database 200 and compare it with the specific value X calculated by the comparison determination unit 340. It is determined whether the value X matches or falls within a predetermined range stored with the reference value (step S106). At this time, the comparison / determination unit 340 determines the age and gender of the reference value that matches the specific value X or the reference value belonging to a predetermined range in which the specific value X is stored together with the reference value. Is determined.

  Finally, the overall control unit 370 controls the display control unit 360 to output the determination result to the display device 400 (step S107) and terminates this operation.

(Judgment process (II))
Next, a determination process for determining the presence or absence of skin abnormality including a calculation process for calculating a fractal dimension based on skin image data acquired by imaging in the present embodiment will be described with reference to FIG. FIG. 8 is a flowchart showing a determination process for determining the presence or absence of skin abnormality including a calculation process for calculating a fractal dimension in the present embodiment.

  In this operation, it is assumed that the standard fractal dimension for each age and gender used when the age of the subject is a determination target is already stored in the database 200 in advance. In addition, it is assumed that a predetermined region of the skin of the subject has already been captured by the imaging camera device 100 and skin image data has been acquired.

  First, when it is set by the user operation of the operation unit 500 that the presence or absence of the subject's skin abnormality is determined as a human characteristic of the subject (step S201), the overall control unit 370 is operated by the display device 400 and the operation unit 500. The subject's sex and age are input (step S202).

  Next, the overall control unit 370 causes the division processing unit 320 to change the number of divisions to generate a plurality of divided images from 2 divisions to n divisions of the original image formed by the skin image data (step S203). For example, as described above, the division processing unit 320 generates respective divided images from 2 divisions to 8 divisions.

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate a fractal dimension in each partial image in the original image and each divided image, and calculates a representative value V of the fractal dimension for each number of data and for each division number ( Step S204).

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate a fractal dimension function with the number of data as a variable for each division number based on the representative value V of the fractal dimension calculated for each number of data and for each division number. Generate (step S205).

  Next, the overall control unit 370 causes the fractal dimension calculation unit 330 to calculate the intersection point on the graph in each function as the specific value X based on the fractal dimension function of each division number (step S206). At this time, when the functions of the fractal dimension do not intersect at one point, the function having the maximum number of functions at the intersection is calculated as the specific value X.

  Next, the overall control unit 370 causes the comparison / determination unit 340 to read out the reference value of the fractal dimension that matches the age and gender input from the database 200 in step SS201 and compare it with the specific value X calculated by the comparison / determination unit 340. Then, it is determined whether the specific value X coincides with or falls within a predetermined range stored together with the reference value (step S207). At this time, the comparison determination unit 340 determines that there is no abnormality in the subject's skin when the specific value X matches or falls within a predetermined range stored together with the reference value. When the specific value X does not belong to the predetermined range stored together with the reference value, it is determined that the subject's skin is abnormal.

  Finally, the overall control unit 370 controls the display control unit 360 to output the determination result to the display device 400 (step S208) and terminates this operation.

[Operation and Effect of First Embodiment]
With the above configuration, the skin feature determination system S of the present embodiment can extract skin features as quantitative feature amounts relatively easily by calculating the fractal dimension in an image of a predetermined region of the subject's skin. The fractal dimension of the skin image can be accurately determined while eliminating the effects of statistically existing artifacts on the skin image using a plurality of original images on the skin of the acquired subject and the divided images obtained by dividing the original image. Since it can be specified, it is possible to accurately determine the human body characteristics of the subject based on the skin image such as the age (skin age) of the subject, sex, and presence / absence of abnormal skin condition.

  In addition, the skin feature determination system S of the present embodiment can eliminate artifacts by using a plurality of original images and divided images obtained by dividing the plurality of original images, and calculate an accurate fractal dimension in the skin of the subject. it can.

  In addition, the skin feature determination system S of the present embodiment can calculate fractal dimensions in various skin regions A by acquiring different skin regions A as skin image data. Can be improved.

  Further, the skin feature determination system S of the present embodiment calculates a fractal dimension in which the influence of artifacts existing in the skin area A is eliminated by skin image data acquired by imaging the same skin area A multiple times. Therefore, the accuracy of the specified specific value X can be improved.

  In addition, the skin feature determination system S of the present embodiment can calculate the fractal dimension by performing a three-dimensional analysis of the shape and gray scale in the original image region and the divided image region. The fractal dimensions of the partial images can be calculated accurately.

  Further, the skin feature determination system S of the present embodiment uses the power spectrum method to calculate the fractal dimension in each original image and each partial image of each divided image, thereby using the power spectrum method. Since the calculation load can be reduced by using a simple formula for the three-dimensional analysis of the shape and gray scale, the fractal dimension in each original image and each partial image can be calculated accurately and at high speed.

  Moreover, since the skin feature discrimination system S of this embodiment has an imaging means, each process can be performed by an integrated apparatus from the acquisition of the skin image to the discrimination of the human body P feature. Without performing complicated operations such as connection with the apparatus, it is possible to accurately determine the human characteristics of the subject such as the age (skin age), sex, and presence / absence of an abnormal skin condition.

(Modification)
In this embodiment, the overall control unit 370 performs arithmetic processing or other processing while controlling each unit in accordance with the program stored in the ROM / RAM 380. However, each program is executed using a computer. The fractal dimension of the subject is calculated based on the captured original image (that is, the specific value X is calculated) to determine the human characteristics of the subject, and the determination result is displayed on the display device 400. Also good.

[Second Embodiment]
Next, a second embodiment of the skin feature discrimination system S according to the present invention will be described using FIG. FIG. 10 is a block diagram showing the configuration of the skin feature determination system S in the second embodiment.

  In the present embodiment, in the first embodiment, the database 200 is fixedly installed at a place different from the place where the human feature is discriminated and data is supplied to the feature discriminating apparatus 300 via the public network L. The other features are the same as those of the first embodiment, and the same members are denoted by the same reference numerals and the description thereof is omitted.

  As illustrated in FIG. 10, the determination apparatus 300 according to the present embodiment includes an input interface 310, a division processing unit 320, a fractal dimension calculation unit 330, a comparison determination unit 340, an operation unit 500, and a display control unit 360. , A camera control unit 350, a communication control unit 390, a ROM / RAM 380, and an overall control unit 370.

  In addition, the communication control unit 390 establishes a communication line with the database 200 that is fixedly installed at a position separated from the determination device 300 in a wired or wireless manner via a public network line L such as a telephone public network or the Internet, and exchanges data. Is supposed to do.

  As described above, in addition to the same effects as those of the first embodiment, the skin feature determination system S of the present embodiment uses the database 200 that is fixedly installed remotely from the place where the image data of the subject is acquired, and the human characteristics of the subject. Since the discrimination can be performed and the reference value can be shared with the other skin feature discrimination system S, it can be realized easily and at low cost, and a plurality of reference values in the database 200 can be updated. In the skin feature discrimination system S, the feature discrimination of the human body P can be performed in the latest state.

A ... Skin region S ... Skin feature discrimination system V ... Representative value X ... Specific value 100 ... Imaging camera device 200 ... Database 300 ... Discrimination device 310 ... Input interface 320 ... Division processing unit 330 ... Fractal dimension calculation unit 340 ... Comparison judgment unit 350 ... Camera control unit 360 ... Display control unit 370 ... Overall control unit 380 ... ROM / RAM
390 ... Communication control unit 400 ... Display device 500 ... Operation unit

Claims (10)

Acquisition means for acquiring a plurality of images of a predetermined region on the skin of the subject as skin image data;
A dividing unit that divides an image area of each original image formed by the acquired skin image data into a plurality of parts and generates a divided image formed from a plurality of partial images;
Fractal dimension calculation means for calculating the fractal dimension of each original image and each partial image;
First, the representative value of the fractal dimension in the original image is calculated as the first representative value based on the calculated fractal dimension in each original image for each number of data. 1 representative value calculating means;
The number of data of the divided image used for the calculation is changed, and for each number of data, the representative value of the fractal dimension in the divided image is set as the second representative value based on the calculated fractal dimension in each partial image for each of the divided numbers. Second representative value calculating means for calculating
Based on the calculated first representative values and the calculated second representative values for each of the division numbers, the first function, the number of data, and the second data using the number of data and the first representative value as variables. Function calculating means for calculating each second function for each of the number of divisions using a representative value as a variable;
Specifying means for specifying one fractal dimension determined based on the calculated first function and each of the second functions as a specific value;
A comparison means for comparing the identified specific value with a fractal dimension as a reference value stored in advance;
Discriminating means for discriminating human characteristics of the subject who acquired the skin image data based on the comparison result;
A skin feature determination system characterized by comprising: The skin feature determination system according to claim 1,
A skin feature discrimination system, wherein the discrimination means discriminates a human characteristic of a subject who acquired the skin image data of at least one of age and sex based on the comparison result. The skin feature determination system according to claim 1,
The skin feature discrimination system, wherein the discrimination means discriminates whether or not the skin is in a normal state as a human body feature of a subject who has acquired the skin image data based on the comparison result. In the skin feature discrimination system according to any one of claims 1 to 3,
The skin feature determination system, wherein the dividing unit generates a plurality of divided images obtained by dividing each of the original images into a plurality of different numbers of divisions by changing the size of a divided region used for the division. In the skin feature discrimination system according to any one of claims 1 to 4,
A skin feature determination system, wherein the acquisition unit acquires a plurality of skin images of different skin regions as skin image data. In the skin feature discrimination system according to any one of claims 1 to 4,
A skin feature determination system, wherein the acquisition means acquires a plurality of skin images of the same skin region as skin image data. In the skin feature discrimination system according to any one of claims 1 to 6,
The skin feature determination, wherein the fractal dimension calculating means calculates a fractal dimension of each of the original image and each of the partial images based on a shape and a gray scale in the original image region and the divided image region. system. In the skin characteristic discrimination system according to claim 7,
The skin feature discriminating system, wherein the fractal dimension calculating means calculates a fractal dimension of each original image and each partial image by the power spectrum method. An acquisition step of acquiring a plurality of images of a predetermined region on the skin of the subject as skin image data;
A dividing step of dividing the image area of each original image formed by the acquired skin image data into a plurality of parts and generating a divided image formed from a plurality of partial images;
A fractal dimension calculating step for calculating a fractal dimension of each of the original images and the partial images;
First, the representative value of the fractal dimension in the original image is calculated as the first representative value based on the calculated fractal dimension in each original image for each number of data. 1 representative value calculation step;
The number of data of the divided image used for the calculation is changed, and for each number of data, the representative value of the fractal dimension in the divided image is set as the second representative value based on the calculated fractal dimension in each partial image for each of the divided numbers. A second representative value calculation step to calculate as
Based on the calculated first representative values and the calculated second representative values for each of the division numbers, the first function, the number of data, and the second data using the number of data and the first representative value as variables. A function calculation step of calculating each second function for each of the number of divisions with a representative value as a variable;
A specifying step of specifying one fractal dimension determined based on the calculated first function and each second function as a specific value;
A comparison step of comparing the specified specific value with a fractal dimension as a reference value stored in advance;
A determination step of determining the human body characteristics of the subject who acquired the skin image data based on the comparison result;
A skin feature discrimination method characterized by comprising: Computer
A skin feature discrimination program for discriminating human characteristics of a subject based on skin image data from a plurality of images obtained by imaging a predetermined area on the subject's skin,
The computer,
A dividing unit that divides an image area of each original image formed by the acquired skin image data into a plurality of parts and generates a divided image formed from a plurality of partial images;
Fractal dimension calculating means for calculating the fractal dimension of each original image and each partial image;
First, the representative value of the fractal dimension in the original image is calculated as the first representative value based on the calculated fractal dimension in each original image for each number of data. 1 representative value calculation means,
The number of data of the divided image used for the calculation is changed, and for each number of data, the representative value of the fractal dimension in the divided image is set as the second representative value based on the calculated fractal dimension in each partial image for each of the divided numbers. Second representative value calculating means for calculating as
Based on the calculated first representative values and the calculated second representative values for each of the division numbers, the first function, the number of data, and the second data using the number of data and the first representative value as variables. Function calculating means for calculating each second function for each of the number of divisions using a representative value as a variable;
A specifying means for specifying one fractal dimension determined based on the calculated first function and each of the second functions as a specific value;
A comparing means for comparing the specified specific value with a fractal dimension as a reference value stored in advance;
Discriminating means for discriminating human characteristics of the subject who acquired the skin image data based on the comparison result,
Skin feature discrimination program characterized by functioning as

Images