JP2014180284A - Skin image generation system for analysis and skin image generation method for analysis - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a skin image generation system for analysis capable of obtaining the image data in which at least texture is mainly photographed and the image data in which stain is mainly photographed, from the image data acquired by a single photographing with one kind of illumination.SOLUTION: A skin illuminated with a white LED12 is photographed with a camera 2 attached with a conversion lens 11. Image data is obtained such as a difference between the image data of red component and the image data of blue component, in the image data obtained by photographing. At that time, brightness levels are matched each other between the blue component and the red component and further, magnitudes of amplitude of waveform are matched each other, before obtaining a difference. The blue component contains texture image and stain image of a skin surface, and the red component mainly contains texture image of the skin surface. Therefore, the image data of the difference contains mainly the image of stain. The image data of the difference is used for analyzing stain. The image data of the red component is used for analyzing texture.

Description

  The present invention relates to an analysis skin image generation system and an analysis skin image generation method for generating an image for analyzing skin texture and skin spots.

Conventionally, the skin condition is analyzed from image data obtained by photographing the skin.
For example, skin texture, skin spots, skin color, and the like are analyzed from the image data. Skin texture refers to the condition of the skin surface. The skin surface has a number of grooves (skin grooves) that cross each other, and the skin is a skin that is slightly raised between these grooves. ing. Skin has pores.

  For example, when the texture is good, the pores are not conspicuous, the skin groove is shallow, and the state is relatively orderly. Such skin texture indicates the state of the skin surface, and in order to analyze the texture, an image of the skin surface is used. The skin color can also be analyzed using a color image obtained by photographing the skin surface.

  On the other hand, since the stain is formed by depositing melanin slightly below the skin surface, it is necessary to photograph a layer slightly inside the skin surface. When the skin is illuminated and photographed, not only the illumination light is reflected on the skin surface, but also a part of the illumination light is transmitted through the skin surface and a spot in the skin surface layer is photographed. However, in the image photographed in this way, an image by reflected light on the skin surface is mainly photographed, and the spot image is hidden by the image on the skin surface, and this spot is not clearly visible. ing.

  Therefore, an image based on the reflected light that has been transmitted through the skin surface and then reflected in the skin surface layer using two polarizing plates to make it difficult to capture an image based on the reflected light from the skin surface. It is done to be in a state of being photographed.

  That is, as shown in FIG. 8, polarizing plates 43 and 44 are arranged on the illumination light source 41 side and the photographing lens (camera 42) side, respectively, and the polarization directions d1 and d2 of these polarizing plates 43 and 44 are set. The skin is photographed by arranging them so as to be orthogonal to each other. In such a case, the reflected light on the skin surface 45 does not change the polarization direction d1 polarized by the light source side polarizing plate 43, and the polarization direction d2 in which the polarization direction on the lens side is orthogonal to the polarization direction d1. Cannot pass through the polarizing plate 44. On the other hand, the light reflected after being transmitted into the skin surface layer portion 46 has a polarization direction random due to multiple scattering of light within the skin surface layer portion 46, and the polarization direction d2 of the polarizing plate 44 on the lens side. The light passes through the polarizing plate 44 because it contains a large amount of light in a polarization direction that is not orthogonal to the polarizing plate. Thereby, the image of the spot in the skin surface layer part 46 from which the reflected light of the skin surface 45 was cut can be obtained (for example, refer patent document 1).

  For example, a white LED is used as the light source 41 described above. There are a plurality of types of white LEDs. For example, a white LED including a blue LED that emits blue light and a phosphor that emits yellow fluorescence when irradiated with blue light can be used as the light source 41. . As shown in FIGS. 9 and 10, the light source spectrum of such a white LED has a narrow peak of the blue wavelength based on the blue LED and a peak at the yellow wavelength based on the phosphor, and from green to red. It consists of wide peaks.

  Here, the white LED (white LED in FIG. 10) whose light source spectrum is shown in the graph of FIG. 10 is relatively compared with the white LED (white LED in FIG. 9) whose light source spectrum is shown in the graph of FIG. The yellow peak is higher than the blue peak. Accordingly, the white LED in FIG. 10 has a high skin color rendering property and is suitable for photographing skin texture. On the other hand, the white LED in FIG. 9 has a blue light peak higher than a yellow light peak and does not necessarily have high skin color rendering. However, the melanin constituting the stain has a higher absorbance for light having a shorter wavelength than light having a longer wavelength, for example, absorbance for blue light is higher than that for red light.

  Accordingly, when shooting a spot, the peak of light of blue wavelength is relatively higher than that of light of shorter wavelength than the white LED of FIG. When the white LED is used, the difference in brightness between the portion where the dark melanin appears darker and the portion where the dark melanin appears brighter becomes larger and the image of the stain becomes clearer.

  Therefore, when photographing the skin, both the white LED of FIG. 9 and the white LED of FIG. 10 are prepared, the white LED of FIG. 9 is used for photographing the skin stain, and FIG. 10 is used for photographing the skin texture. It is preferable to use two types of illumination (white LED) properly so as to use white LEDs.

JP 2008-289706 A

  By the way, when analyzing skin texture, skin spots, etc. from the captured image data, when taking an image for analyzing skin spots using two polarizing plates as described above, Since the skin surface image is hardly included, it is necessary to take an image for analyzing skin texture separately from the skin stain image.

  That is, when analyzing at least both skin texture and skin stain, an image for spot analysis taken using two polarizing plates and a texture analysis taken without using a polarizing plate An image is required, and two images are required for one analysis.

Further, as described above, it is preferable to use illumination light when photographing skin texture and illumination light when photographing skin spots, and two types of LEDs are used as illumination LEDs, and the skin The illumination is switched between photographing the texture and photographing the skin stain, and the photographing becomes complicated together with the switching of the presence or absence of the polarizing plate on the illumination side.
For the analysis of the skin color, if the image data taken for skin texture is color, this image data can be used.

In addition, there is a demand to easily analyze the skin condition at home, and a skin camera for home use that captures the texture analysis image and the stain analysis image as described above has been proposed.
For example, such a skin camera is connected to a portable electronic device such as a smartphone, and images captured by the skin camera are captured in the smartphone and processed, or the captured image data is transmitted from the portable electronic device via the Internet. It has been proposed to transmit to a server that provides services such as skin image analysis.
In addition, a lens module having a conversion lens that is mounted at a position corresponding to the camera portion of a smartphone and that can capture an image for skin analysis with the camera of the smartphone has been proposed.

  In the lens module having the skin camera or the conversion lens, in order to capture the above-described spot analysis image and texture analysis image, for example, a spot photographing illumination device that illuminates via a polarizing plate, and a polarizing plate It is necessary to provide a texture photographing illumination device that illuminates without intervention and to photograph the skin at least twice by switching these illumination devices. That is, an illumination LED including a polarizing plate is required in addition to a normal illumination LED, which increases the cost of a skin camera and a lens module. In this case, for example, a polarizing plate is also arranged on the lens side, but it is not necessary to move the polarizing plate between the texture photographing and the spot photographing.

  Further, in a skin camera, a camera, a lens, and illumination are included in one device, and these can be controlled by a control device of the skin camera. On the other hand, when the lens module is attached to the camera of the smartphone, the LED lighting of the camera provided on the smartphone is set to shoot a subject about 1 m away from the camera. When photographing skin, there is a possibility that the color will be too bright and white. Moreover, the LED illumination provided in the smart phone is illumination for normal photography, and there is no polarizing plate.

  Therefore, it is necessary to provide a lighting device for skin photographing on the lens module side. For example, when a lens module is provided with an illumination LED for skin spot photography with a polarizing plate and an illumination LED for skin texture photography without a polarizing plate, a lens is required because a polarizing plate is required. Module costs increase.

  In addition, it is necessary to switch between the two types of lighting to shoot, but when the lens module and the smartphone are not connected to each other so that data communication is possible, the two types of lighting are automatically switched to shoot twice. It is difficult to control. When shooting twice with manual illumination switching, there is a risk that the user may find it cumbersome to manually switch between the two types of illumination and perform twice imaging.

  If the lens module is equipped with various known communication means such as USB, Bluetooth (registered trademark), WiFi, NFC, etc. that can communicate with the smartphone so that it can automatically switch between the two types of illumination and shoot, the lens Module costs increase.

  The present invention has been made in view of the above circumstances, and image data in which at least mainly texture (skin, skin, pores, etc.) is captured from image data obtained by one shooting with one kind of illumination. Another object of the present invention is to provide an analysis skin image generation system and an analysis skin image generation method capable of obtaining image data in which spots are mainly reflected.

In order to solve the above-described problem, the skin image generation system for analysis of the present invention is a skin image generation system for analysis that generates skin image data that can analyze skin spots,
Illumination means capable of irradiating the subject with illumination light including at least red light and blue light;
Photographing means for photographing the subject illuminated by the illumination means;
Storage means for storing image data of the subject photographed by the photographing means;
Image processing means for processing image data of the subject stored in the storage means,
The image processing means is an image extracting means for extracting red component image data and blue component image data from the image data stored in the storage means;
The difference image which is the difference between the blue component image data and the red component image data by matching the brightness levels of the red component image data and the blue component image data extracted by the extraction means And a difference calculating means for calculating data.

The skin image generation method for analysis of the present invention is a skin image generation method for analysis that generates skin image data that can analyze skin spots,
Extracting red component image data and blue component image data from image data taken by illuminating the subject with illumination light including at least red light and blue light,
After matching the brightness levels of the extracted red component image data and the blue component image data, difference image data that is the difference between the blue component image data and the red component image data is obtained. It is characterized by seeking.

  According to such a configuration, in order to analyze the skin texture from one image data obtained by photographing the skin as a subject illuminated with illumination light including at least blue light and red light. Image data for analyzing the texture of the skin where the surface of the skin is photographed and image data for analyzing the spot of the skin where the spots in the surface layer of the skin are photographed (if there are spots) can be obtained.

  A spot made of melanin absorbs blue light more than red light. Therefore, it is possible to photograph the spot more clearly when shooting with blue light than when shooting with red light. In other words, the spot image becomes thin in an image taken with red light, and it becomes difficult to analyze the spot. In addition, in an image photographed with red light, since the stain becomes thin, the image on the skin surface becomes clearer.

  Blue wavelength light is not necessarily highly permeable to the inside of the skin, but blue light is transmitted to the skin surface layer where the spots are formed, that is, to the shallow part of the skin, and to melanin as described above. Photographed so that spots are darkened after being absorbed. However, since the blue reflected light reflected on the skin surface is photographed, the skin surface image and the spot image are mixed, and the spot image cannot be clearly recognized.

  Therefore, when the red component image data is subtracted from the blue component image data, the skin surface image is erased from the blue component image data, and a stain image (if there is a stain) in the skin surface layer portion remains.

  However, since the red component image data and the blue component image data have different levels of brightness (luminance) as a base, before calculating the image data that is the difference between the red component image data and the blue component image data. Therefore, it is necessary to match the brightness level between the red component image data and the blue component image data.

  The difference image data obtained by subtracting the red component image data mainly including the skin surface image from the blue component image data includes an image suitable for analyzing a stain in the skin surface layer portion. Become. As the image data for analyzing the texture, the image data of the subject that is illuminated by the illumination unit and photographed by the photographing unit can be used as it is, and the red component image data that does not include a spot image as described above. Can be used.

  In this case, the red component image data, which is mainly the skin surface image (image without the stain), is subtracted from the blue component image data that includes both the stain and skin surface images to a certain level or more. Thus, since the image of the skin surface can be removed, it is not necessary to use two polarizing plates for photographing a spot as in the prior art, and the costs of the photographing means and the illuminating means can be reduced.

  In addition, by performing one image with one type of illumination, both image data for texture analysis and image data for stain analysis can be obtained, so illumination that passes through the polarizing plate as in the past. And it is not necessary to switch between two types of illumination, that is, illumination that does not pass through, and photograph twice.

  Therefore, when shooting with a lens module having a conversion lens in the camera of a smartphone, it is not necessary to switch between the two types of illumination and shoot twice. Therefore, it is not necessary to use a relatively expensive communication means between the lens module and the smartphone. Therefore, it is easy to photograph the texture and the stain, and the costs of the photographing means and the illumination means can be reduced.

  Further, the illumination light including the red wavelength and the blue wavelength may be white light, and the image data including the red, blue, and green components is converted into skin light by using the illumination light of the illumination unit as white light. It can be image data for analyzing colors. Accordingly, it is possible to obtain three types of image data (two types when the texture and color are analyzed with the same image data) that can analyze the texture, stain, and color of the skin with one shooting.

  Further, when the image data is stored even after the analysis of the skin condition, if the original image data at the time of shooting is stored, one image data may be stored for one shooting of one person. Therefore, for example, when storing and storing image data sent from many users in a skin image analysis service via the Internet, the storage capacity can be reduced.

  In the analysis skin image generation system and the analysis skin image generation method of the present invention, the difference image data is used as image data for analysis of skin spots, and the red component image data is used for analysis of skin texture. It is preferable to use image data.

  According to such a configuration, the difference image data calculated from one image data obtained by photographing the skin can be used for analysis of skin spots, and the difference image data is extracted from the image data at the time of calculation. The red component image data can be used for analysis of skin texture. Therefore, it is not necessary to change the presence or absence of the polarizing plate in the illumination light as in the past, and to perform shooting for skin spots and skin texture separately, and image data for analyzing spots and textures. Can be easily acquired.

  In the skin image generation system for analysis according to the present invention, the difference calculation means stage sets the amplitudes of waveforms that approximate each other in the red component image data and the blue component image data extracted by the extraction means to each other. It is preferable to obtain the difference image data after matching them so that they are substantially the same.

  Further, in the analysis skin image generation method of the present invention, before obtaining the difference image data, the amplitudes of the waveforms that are approximate to each other in the extracted red component image data and the blue component image data are mutually equal. It is preferable to match them so that they are substantially the same.

  According to such a configuration, the red component image data and the blue component image data each include an image of the skin surface, and since the same skin surface is captured, the horizontal or vertical direction of the image data However, since the brightness level differs between the red component image data and the blue component image data as described above, it is necessary to match the brightness levels. Furthermore, even if the waveform based on the skin surface image of the blue component image data and the red component image data is approximate, the red component and the blue component have similar amplitudes, so when obtaining the difference image data, By adjusting not only the brightness level but also the amplitude, the image of the skin surface can be more accurately removed from the blue component image data.

Further, in the analysis skin image generation system of the present invention, the photographing means comprises a camera mounted on a portable electronic device and a skin photographing conversion lens provided in a lens module attached to the camera. And
It is preferable that the illumination unit is provided in the lens module.

  According to such a configuration, since it is possible to obtain image data that can analyze both the texture and the stain by photographing once with one type of illumination without changing the type of illumination and photographing multiple times, the lens Skin can be easily photographed without providing a communication means between the module and a portable electronic device such as a smartphone.

  According to the present invention, it is possible to analyze the texture and the stain by one shooting under one kind of illumination, and the shooting can be facilitated. In addition, a polarizing plate is not required for spot photography, and the cost can be reduced.

It is a schematic sectional drawing which shows the smart phone with which the lens module of the skin image generation system for analysis which concerns on embodiment of this invention was mounted | worn. It is a schematic front view which shows the conversion lens in the said lens module, and a circuit board provided with LED. It is a block diagram which shows the outline of the skin image generation system for analysis. It is a block diagram which shows the outline of the modification of the skin image generation system for analysis. It is a figure for demonstrating the calculation method of the image data for skin spot analysis. It is a figure for demonstrating the calculation result of the image data for skin spot analysis, (a) is a figure of the image data which displayed the image of the image | photographed color in gray scale, (b) The image | photographed color FIG. 6C is a diagram of image data in which the red component image data extracted from the image data is displayed in gray scale, and (c) is a diagram of the blue component image data and the red component image data extracted from the color image data. It is the figure of the image data which showed the image data used as a difference by the gray scale. It is a several graph for demonstrating the calculation method of the image data for skin spot analysis. It is a figure for demonstrating the imaging | photography principle of the skin spot image using a polarizing plate. It is a graph which shows the light source spectrum of white LED suitable for a skin spot image photography. It is a graph which shows the light source spectrum of white LED suitable for the texture image photography of skin.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
As shown in FIGS. 1 and 2, the analysis skin image generation system of this embodiment includes a smartphone 1 as a portable electronic device including a digital camera (a camera 2 including an imaging device), and a camera of the smartphone 1. 2 is provided with a conversion lens 11 composed of two lenses 11a and 11b for enlarging the skin and taking a close-up image, and a lens module 10 having an LED 12 for photographing illumination.

  In the skin image generation system for analysis, the above-described LED 12 serves as an illumination unit that illuminates the skin, and the camera 2 and the conversion lens 11 serve as an imaging unit that photographs the skin illuminated by the illumination unit. Further, the analysis skin image generation system includes storage means for storing image data obtained by photographing the skin as a subject illuminated by the illumination means with the photographing means, and image data for spot analysis from the image data, Image processing means for generating image data for texture analysis. This storage means may be a later-described flash memory 4 of the smartphone 1, and the image processing means may be a later-described control device 3 of the smartphone 1.

  That is, as shown in FIG. 3, the smartphone 1 is mounted with a lens module 10 that includes a conversion lens 11 and an illuminating unit a that irradiates at least blue light and red light at the position of the camera 2. The conversion lens 11 and the camera 2 constitute a photographing means b. Further, the smartphone 1 includes a storage unit c configured by the flash memory 4, and skin image data illuminated by the illumination unit a and photographed by the photographing unit b is temporarily stored in the storage unit c.

  In addition, the smartphone 1 includes a communication unit d and can be connected to the Internet h by connecting to a mobile communication network or connecting to a wireless LAN, and the skin image stored in the storage unit c described above. Data is transmitted to a data management server (server) g via the Internet h. The data management server g analyzes the skin state from the skin image data transmitted from each smartphone 1, and determines how good (or bad) the skin state is.

  In addition, the data management server g manages skin image data, data of skin condition analysis determination results, and the like for each user, for example, and enables the user to view these data using the smartphone 1. Data such as skin image data and judgment results of skin condition analysis can be used by excluding information that can identify individuals from companies related to skin such as a company.

  Such a data management server g includes a communication means i that enables data communication with the smartphone 1 via the Internet h, skin image data transmitted from the smartphone 1 via the Internet, and analysis of the skin image data. The skin history database j in which the skin condition determination result and the like are registered is provided. The data management server g also includes a control device that functions as the image processing means k, and this control device extracts blue component image data and red component image data from the captured image data, as will be described later. Functions as the extraction means e.

  Further, the image processing means k adjusts the brightness level between the blue component image data and the red component image data as described later, and further, the image signal constituting the red component image data and the blue component image Functions as difference calculation means f for calculating difference image data that is the difference between the blue component image data and the red component image data after matching the magnitudes of the amplitudes of the waveforms approximated by the image signals constituting the data. .

  The image processing unit k uses the above-described red component image data, generates a determination result of the quality of the skin texture using a known determination algorithm, and uses the above-described difference image data to generate a known result. By the determination algorithm, it functions as an analysis determination unit m that generates a determination result of whether the skin stain is good or bad. The determination result by the analysis determination unit m is stored in the skin history database j as a storage unit in association with each user.

  As shown in FIG. 4, the processed skin image data may be processed by the smartphone 1, and the smartphone 1 described above is in addition to the illumination unit a, the imaging unit b, and the storage unit c described above. Image processing means k, extraction means e, difference calculation means f, and analysis determination means m. The data of the determination result by the analysis determination unit m is stored in the storage unit c in association with the skin image data used for the analysis.

  Further, the photographing means b is not only the camera 2 and the conversion lens 11 of the lens module 10 described above, but is a single skin camera 5 capable of photographing the skin as in the case of using the conversion lens 11. Also good. The skin camera 5 is capable of data communication with the smartphone 1 via a communication means such as USB, for example, and the captured image data is transmitted from the skin camera to the smartphone 1 and stored in the storage means c. The Note that a storage device such as a flash memory may be provided on the skin camera side so that image data captured by the skin camera can be stored.

  The lens module 10 includes the conversion lens 11 and the LED 12 described above, an LED drive circuit 14 that drives the LED 12, and a housing 10a that accommodates these. Within the housing 10a, the two lenses 11a and 11b of the conversion lens 11 are supported by the barrel 20a, and in front of the master lens (not shown) of the camera 2, the optical axis of the master lens of the camera 2 and the two lenses It arrange | positions so that the optical axis of lenses 11a and 11b may be matched.

  The housing 10a includes a lens housing 20 on the front side (subject side) of the barrel 20a. The lens housing 20 has a front end surface that is a contact portion 20b that comes into contact with the skin when photographing the skin as a subject, and blocks outside light in a state where the contact portion 20b is in contact with the skin. Further, an opening 20c is formed in the contact portion 20b, and a portion facing the skin opening 20c is photographed.

  The barrel 20a is fixed to the base end side (the side attached to the camera 2) of the lens housing 20 in the housing 10a while supporting the conversion lens 11 including the lens 11a and the lens 11b. Further, in the housing 10a, as shown in FIGS. 1 and 2, the LED 12 that irradiates photographing illumination light toward the opening 20c of the contact portion 20b of the lens housing 20 and the LED 12 are driven. A circuit board 13 having an LED drive circuit 14 is supported.

  Further, for example, one LED 12 is provided on the circuit board 13, but for example, a plurality of LEDs 12 may be provided so as to illuminate the skin from different positions.

  In order to shoot an image with a clear skin texture, it is preferable to arrange one LED 12 that irradiates light obliquely to the skin, but for averaging the brightness of the skin shooting range, It is preferable to arrange a plurality of LEDs 12 at a plurality of locations. Also, the texture is basically skin irregularities, and by irradiating light obliquely onto the skin from one light source, the shadows of the skin irregularities become clear and the image of the skin irregularities becomes easy to understand.

Since the LED drive circuit 14 is provided with a booster circuit and a constant voltage circuit, the LED illuminance lowering problem during use does not occur.
Further, the lens housing 20 is in a light shielding state so that ambient light does not enter the imaging region. Accordingly, a substantially constant shooting condition can be secured at all times, and an image correction process based on a difference in shooting condition for each shooting is not required.

The lens module 10 also includes a power supply circuit 18 including a battery that supplies power to the LED 12, the LED drive circuit 14, and the like, and a power switch (not shown).
The power switch may be a mechanical switch that is manually turned on / off.

Further, a structure may be adopted in which a switch is automatically turned on when the abutting portion 20b of the lens housing 20 abuts against the skin by providing a push switch in the abutting portion 20b of the lens housing 20 against the skin.
For example, the lens housing 20 and the contact portion 20b are separate members, and the contact portion 20b is supported by the lens housing 20 so as to be movable from the lens housing 20 to the subject side, and is elastically (biasing means). The contact portion 20b is biased toward the subject, and the lens housing 20 and the contact portion 20b are always separated from each other by a predetermined distance. The lens housing 20 is provided with a switch that is turned on when a push switch or the like is pressed between the lens case 20 and the contact portion 20b and turned off when the pressure is released. With such a structure, when the contact portion 20b is applied to the skin that is the subject and the contact portion 20b moves toward the lens housing 20 against the urging force of the elastic member, the switch is pressed. When the contact portion 20b is separated from the skin, the contact portion 20b biased toward the skin side is separated from the switch and the power is turned off.
In addition, a photoelectric switch is provided inside the lens housing 20, and when the skin is brought into contact with the contact portion 20 b of the lens housing 20 and the entrance of light into the lens housing 20 is blocked, the inside of the housing is The photoelectric switch may be turned on when dark.

  As is well known, the smartphone 1 has a function as a mobile phone and can make a call using a wireless line. The smartphone 1 has a display (not shown), can be connected to the Internet via a wireless line, and can send and receive e-mails via the Internet. Furthermore, the browser of an application (application) enables browsing of a website. In addition, for example, a file can be uploaded to a server such as a website, and the file can be downloaded from the server such as a website.

The smartphone 1 includes a control device 3 having a CPU, a ROM, a RAM, and the like, and can execute an application as the above-described program. The control device 3 is connected to a flash memory 4 as a storage device, and can store downloaded applications (programs), music files, video files, and the like. In other words, the moving image file and the still image file include image data captured by the camera 2.
Further, the control device 3 can take a picture by controlling the camera 2 and can save the taken image data in the flash memory 4. Further, image data analysis and analysis processing can be performed by an application for image processing and image analysis.

In this embodiment, image data that can analyze a spot is generated by image processing still image data obtained by photographing the skin as a subject illuminated by the LED 12 with the camera 2 using the lens module 10. The image processing is performed by the control device 3 of the smartphone 1.
As described above, image data is sent from the smartphone 1 to the above-described data management server, and on the data management server side, the image data photographed by the smartphone 1 is stored and image processing of this image data is performed. Image data used for analysis may be obtained.

  Here, in a general digital camera including a camera built in a smartphone or the like, a color filter in which red, green, and blue single color filters are arranged in a predetermined pattern corresponding to each pixel may be used in an image sensor. . In this case, depending on the sensor of each pixel, there are a sensor for photographing the red component, a sensor for photographing the green component, and a sensor for photographing the blue component. The values from the sensors of the respective colors at these positions indicate the brightness (luminance) of the light transmitted through the corresponding filter.

  The luminance data for each pixel may be output as RAW data as it is, but usually a color image having red, green, blue, that is, RGB values for each pixel by interpolation processing, color adjustment, or the like. It is converted into data and output from the camera. Note that the color filter is not necessarily a red, green, and blue pattern, and these complementary colors may be used, or a four-color filter may be used instead of three colors, and the image data is not limited to the RGB system. Although there are other types of image data, in any case, it is possible to convert the color of each pixel into RGB type image data indicated by RGB values.

  Each pixel is image data represented by RGB values, and image data using only the R value for each pixel, that is, image data with G and B values of 0 is red component image data, In the pixel, the R and B values are 0, and the image data using only the G value is the green component image data. In each pixel, the R and G values are 0 and only the B value is used. The image data is blue component image data. That is, in the image data composed of the RGB values of each pixel, by setting two values of RGB to 0, it is possible to extract red component image data, green component image data, and blue component image data.

  When red component image data is displayed, it becomes a red shade image, when green component image data is displayed, it becomes a green shade image, and when blue component image data is displayed, it becomes a blue shade image. In this case, for example, when image data of red component is used for texture analysis, it may be difficult to see an image composed of red and light.

  Therefore, in the red component image data, the G value and the B value can be converted to a black and white image (grayscale image) by setting the value of G and B to the same value as R instead of 0. Similarly, the image data of the green component and the image data of the blue component can be converted into a gray scale image by aligning the RGB values to the G value or the B value.

In addition, image data, which is a difference described later, basically has a luminance difference between a red component and a blue component, and is handled as grayscale image data.
Here, the grayscale image data may be converted to a skin color by image processing.

  In this case, for example, after color-correcting the skin image data photographed as described later, a reference skin color is determined using the corrected skin image, and each RGB value of each pixel is set, A skin color-based texture image / stain image can be obtained.

  For example, RGB values at a predetermined brightness are set. In this case, for example, RGB values are set as R: 98/256, G: 68/256, B: 58/256, etc., and this is used as a skin color pattern and converted to the above skin color. Corresponding to the change in luminance of the image of the grace case, image data composed of a gradation of skin color can be obtained by raising and lowering each RGB value while maintaining the above pattern.

  Further, by calculating a “fluctuation” pattern of pixel luminance values from the skin image and setting each value of RGB, a skin color-based texture image / stain image that seems to be more actual skin can be obtained. That is, a more natural skin color can be expressed by shifting the above-mentioned pattern slightly in each pixel.

  The LED 12 is an illumination means for illuminating the skin, and for example, a white LED is used. There are several types of white LEDs. For example, a white LED that combines a blue LED and a phosphor that glows yellow when irradiated with blue light, and a red, green when ultraviolet light is emitted to an ultraviolet LED that outputs ultraviolet light. There are a white LED that combines a plurality of phosphors that emit blue light, a blue LED that emits blue light, a green LED that emits green light, and a white LED that combines a red LED that emits red light.

  Any of these white LEDs may be used, but the illumination means in this embodiment preferably emits light including at least blue light and red light. The blue light is, for example, light having a wavelength of 450 to 495 nm, and the red light is, for example, light having a wavelength of 620 to 750 nm, and is included in the light emitted from each of the white LEDs described above. In addition, it is not necessary to include all light having wavelengths within the above-described wavelength ranges, and it is only necessary to include some wavelengths in each wavelength range. Moreover, the range of the wavelength of visible light is, for example, 380 to 750 nm, and basically it is sufficient that light on the long wavelength side and the short wavelength side of visible light is included.

Next, image processing performed by the control device 3 of the smartphone 1 as the above-described image processing means will be described.
Here, the stain is formed from melanin deposited in the skin surface layer. In the absorption spectrum of melanin, the shorter the wavelength, the higher the absorbance, and blue light having a shorter wavelength is more absorbed than red light having a longer wavelength.

Therefore, a portion having a large amount of melanin becomes darker in the blue component of the captured image data, but is not so dark in the red component, and the blue component has a clearer spot than the red component.
That is, in the skin surface layer portion, there are a portion of a stain with a large amount of melanin and a portion other than a stain with a small amount of melanin, and in the red component image data, the absorbance of melanin red light is low. Even in many spots, the red light is reflected to the camera 2 side with little absorption. Therefore, the brightness on the image data does not decrease much even at the spot portion.

  In the portion where the amount of the stain is small, red light is further absorbed, but the difference between the luminance of the portion of the stain on the red component image data and the luminance of the portion other than the stain is small. Therefore, in the red component image data, the spot image is a low-contrast image with a small difference between light and dark, and the spot is visible, but there is little difference in brightness between the spot and the part other than the spot and the spot is difficult to see. It becomes image data.

  On the other hand, in the blue component image data, the absorption of blue light from melanin is high, so the blue light reflected from the melanin-rich spot in the skin surface is absorbed by the spot and the camera is absorbed. Reflected to the second side. Accordingly, in the blue image data obtained by photographing the reflection of the blue light that has been absorbed in a large amount by melanin, a stain having a large amount of melanin appears darker. Even in the blue light, the blue light is not so much absorbed in the portion with a small amount of melanin other than the stain. Therefore, the difference between darkness and lightness of the spot portion and the other portions is large, resulting in an image with high contrast, and an image that allows easy viewing of the spot.

  Therefore, although relative to the red component image data, the skin surface image is the main image, and the stain image is considerably thin, whereas the blue component image data is the skin surface image. An image is included, and an image of a darker (darker) spot than that in the case of red is included.

  FIG. 5 is a schematic diagram of image data schematically showing a skin skin groove with a plurality of lines intersecting each other as a skin surface image, and schematically showing a stain in the skin surface layer portion in a cloud shape. The image data of B (blue component) includes an image of the skin surface (skin groove) represented by lines intersecting each other and an image of a stain in the skin surface layer portion represented in a cloud shape (there is a stain) If).

  FIG. 6A shows the captured image data 51 as a grayscale image, and FIG. 6B shows the red component image data 52 as a grayscale image. FIG. 6C shows image data 53 as a difference between the blue component image data and the red component image data 52 in a gray scale as will be described later.

  In addition, the R (red component) image data basically includes the skin surface image and the stain image in the skin surface layer portion, but even if there is a stain image, it is compared with the blue component image data. The stain is thin and the stain is difficult to see. That is, in the image data shown in FIG. 6, the captured image data 51 includes a blue component stain image and is mixed with the texture image. However, in the red component image data 52, the stain image is The image data is thinner and easier to see than the photographed image data.

  Here, the blue component and the red component are, for example, a B value (brightness and luminance) and an R value (brightness and luminance) when image data is expressed in RGB.

  FIG. 7 is a graph (waveform) showing the luminance of each pixel of one horizontal line at the same vertical position on the image data in the red component image data and the blue component image data. 30 shows the waveform of the upper red component and the waveform of the lower blue component. Each of the red component waveform and the blue component waveform includes the same waveform indicating the image of the skin surface schematically shown by a plurality of intersecting lines in FIG. The waveform is similar to the waveform of the blue component.

Further, a part of the difference between the red component waveform and the blue component waveform is based on a waveform due to a stain that is displayed darker than the red component image data in the blue component image data.
Therefore, if the difference between the blue component image data and the red component image data is obtained, the skin surface image is removed from the blue component skin surface image and the stain image, the stain image remains, and the skin surface remains. It is easy to see the spot image without being disturbed by the image.

  However, as shown in the graph 30 of FIG. 7, there is a large luminance difference between the red component image data and the blue component image data. It is necessary to lose.

In addition, the luminance of the red component image data is higher than that of the blue component image data, but when looking at each waveform, the amplitude of the waveform mainly based on the reflected light on the skin surface is the red component image data. The blue component image data becomes larger.
Therefore, in order to obtain differential image data that does not include an image of the skin surface, it is preferable to match the amplitude of the waveform between the blue component and the red component.

  In the graph 30, the baseline of the waveform of the red component image data and the baseline of the waveform of the blue component image data are obtained. The expression in the graph 30 is an expression indicating a baseline. Further, the baseline shown by the above formula is an approximate curve obtained from the waveform of the graph, and here, the approximate curve is obtained as a quadratic curve.

  As indicated by these two baselines, the luminance at the center is high and the luminance at the left and right ends is low. This is because the illumination by the LED 12 is bright at the center of the photographing range of the skin and dark at the periphery, and is uneven in the image due to the illumination. The difference in brightness depending on the position on the image data based on such illumination is reduced by increasing the number of LED installation locations. However, when the skin is illuminated from multiple directions, the shadow disappears. It becomes an image in which some skin grooves and pores are difficult to see.

  That is, the image on the skin surface is image data that is easier to analyze the texture of the unevenness of the skin when the number of light source arrangements is smaller. That is, in the waveform indicating the skin surface, the smaller the number of light source arrangement locations, the larger the amplitude. When the number of light source placement locations increases, the amplitude decreases, resulting in an image in which skin grooves and pores are not conspicuous.

Differences in brightness (unevenness) based on illumination on image data can be corrected by well-known shading correction, and are preferably corrected.
Further, the above-described equation for obtaining the difference image data is, for example, when the luminance level of the red component is R, the luminance level of the blue component is B, and the obtained difference is D, D = α (B−β · (R−k)).

  Here, α, β, and k are coefficients, and k is the difference in the luminance level between the blue component and the red component described above, and the luminance level from the image data of the red component that is higher in luminance than the blue component. By subtracting k as the difference, the level of brightness (brightness) is adjusted between the blue component and the red component. k is obtained from the difference in luminance level between the red component and the blue component, for example, the difference between the above-described baseline of the luminance waveform of the red component and the baseline of the waveform of the luminance of the blue component.

  Β is a coefficient for reducing the difference in amplitude between the red component waveform and the blue component waveform. For example, each peak of the blue component waveform and the corresponding red component of these peaks For example, multiplying the peak height of the red component waveform by β reduces the difference between the red component peak height and the blue component peak height. become.

α is used to adjust the magnitude of the difference data value after the difference is obtained, and is used to adjust the difference value so that the range is easy to analyze.
In RGB image data, each pixel is assigned an RGB value. As described above, image data captured with light illuminated by a white LED including blue light and red light is converted into RGB format. D is obtained by substituting the value of R and the value of B in each pixel into the above-mentioned equation of D = α (B−β · (R−K)).

  In this case, as schematically shown in FIG. 5, by removing the red component image data mainly including skin texture from the blue component image data including the skin texture image and the skin stain image, When the difference image data is obtained, the difference image data is image data mainly having an image of a skin spot. In this image data, as shown in the difference image data 53 in FIG. 6 (c), an image such as a skin groove on the skin surface is removed, and an image mainly showing a stain in the skin surface layer portion is obtained.

  Further, as shown in a graph 30 of FIG. 7, in the red component image data and the blue component image data in which the luminance level and the amplitude of the waveform are different, the luminance level and the amplitude of the waveform are expressed as follows. When combined, as shown in the graph 31, the red component image data and the blue component image data having waveforms approximate to each other overlap each other. However, in these image data, the stains mainly included in the blue component are displayed. There are differences based on images. Therefore, as shown in the graph 32, by obtaining the difference data (D) of these image data, it is possible to obtain difference image data mainly including a spot image without including a skin surface image.

The difference image data composed of the D value obtained for each pixel includes a stain image from the blue component image data including the skin surface image and the stain image in the skin surface layer as described above. By subtracting the image data of the red component that is thin and mainly composed of an image of the skin surface, an image of a stain in the skin surface layer portion that does not substantially include the image of the skin surface is included.
This difference image data becomes image data for analyzing skin spots.

  Further, among the original color image data used for obtaining the difference image data, the red component image data that hardly includes a spot image as described above is used as image data for analyzing the texture of the skin surface. be able to. Note that original color image data that has been affected by a stain but has not been photographed and processed may be used as texture analysis image data.

  The original image data is color image data illuminated by a white LED (LED 12), and this data can be used as image data for skin color analysis. In this case, it is preferable to take a white balance. However, the camera 2 of the smartphone 1 is provided with a white balance function, so that a white balance can be obtained.

  In the analysis skin image generation method in this analysis skin image generation system, the blue component is generated in each pixel of the image data photographed by the white LED including blue light and red light by the function of the above-described image processing means. By calculating the difference between the red component and the red component based on the equation D = α (B−β · (R−K)), it is possible to obtain a spot analysis image. Further, as described above, the original image data or red component image data can be used as image data for analyzing skin texture. Furthermore, image data for skin color analysis can be obtained by white balance of the original image data.

  Here, in order to obtain a color balance such as white balance with high accuracy, for example, when the lens module 10 or the skin camera 5 is used for the first time, that is, once after purchase, the reference white attached to the skin camera is used. The correction value for determining the reference white color of the photographed white image is determined inside the smartphone or on the server side described above, and stored in either the flash memory 4 of the smartphone 1 or a storage measure on the server side. deep.

An appropriate skin image can be obtained by performing color correction on the photographed skin image using the above correction value during or after skin photographing.
In place of the white test chart, a test chart printed with a plurality of or single chromatic colors may be used.

  In addition, the above-described test chart is provided within the photographing range of the lens module 10 and the skin camera 5 so that the test chart is reflected in the photographed image, and a correction value is determined for the reflected test chart every time. You may make it correct the skin picture photoed using

  From the above, in such a skin image generation system for analysis and a skin image generation method for analysis, image data that can analyze the stain on the skin surface layer inside the skin surface without using a polarizing plate is obtained. Since it is possible to obtain the skin, it is not necessary to use polarizing plates whose polarization directions are orthogonal to each other on the photographing illumination side and the photographing lens side in the skin photographing apparatus, and a dedicated skin camera or smartphone for photographing the skin In the lens module 10 having the conversion lens 11 for adding a skin photographing function to the camera of the portable electronic device 1 and the like and the LED 12 for illumination, the cost can be reduced.

  In addition, it is possible to obtain image data for analyzing skin spots, image data for analyzing skin texture, and image data for analyzing skin color by performing a single shooting operation. It is possible to easily shoot an image, and the burden on the user can be reduced when each user shoots skin by himself.

  In addition, when the lens module 10 is attached to a digital camera of a portable electronic device and a skin image is taken, for example, it is possible to use a predetermined illumination without changing the type of illumination or changing various shooting settings. Since the image data that can analyze each of the above-mentioned skin spots, textures, and colors can be obtained by performing the image capturing a number of times, the image capturing can be easily performed without automating the image capturing.

  In addition, even in the case of automating shooting, it is not necessary to switch the illumination and setting and perform shooting a plurality of times as in the prior art, so basically the conversion lens 11 and the lens module 10 side having the LED 12 for illumination are provided. With the power on and the LED 12 lit, for example, the smartphone 1 side activates an application for skin photography to automatically shoot, or conversely, after the application is launched, based on the instructions of the application It is only necessary to automatically take a picture after turning on the power of the lens module 10 to turn on the LED 12, and no particular communication is required between the lens module 10 and the smartphone 1, and the lens module 10 is provided by providing a communication means. It is possible to prevent the cost from increasing.

  In the above-mentioned server, the skin image data transmitted from the smartphone 1 is stored in the skin history database for each user of the smartphone 1, for example, so that the user can view it at any time, or can be used by a cosmetic manufacturer or the like. Can be considered. In this case, it has been necessary to register a texture image and a stain image (+ skin image) so far. On the other hand, in this skin image generation system for analysis, if one piece of skin image data is stored, image data for analysis of texture, stain, and skin color can be obtained at any time. Storage capacity can be reduced.

  The LED 12 is preferably installed at one location so that the shadows of the unevenness are darker in obtaining image data for analyzing the texture that is the unevenness of the skin. Since the surface image is removed, it is preferable to have a plurality of locations in order to eliminate uneven illumination. However, the problem of uneven illumination can be solved by performing the shading correction as described above even if the LEDs 12 are arranged at one location.

1 Smartphone (portable electronic device)
2 Camera (photographing means b)
3 Control device (image processing means k, extraction means e, difference calculation means f)
4 Flash memory (storage means c)
10 Lens module 11 Conversion lens (photographing means b)
11a lens 11b lens 12 LED (illumination means a)

Claims (7)

A skin image generation system for analysis that generates skin image data capable of analyzing skin spots,
Illumination means capable of irradiating the subject with illumination light including at least red light and blue light;
Photographing means for photographing the subject illuminated by the illumination means;
Storage means for storing image data of the subject photographed by the photographing means;
Image processing means for processing image data of the subject stored in the storage means,
The image processing means, extracting means for extracting red component image data and blue component image data from the image data stored in the storage means;
The difference image which is the difference between the blue component image data and the red component image data by matching the brightness levels of the red component image data and the blue component image data extracted by the extraction means A skin image generation system for analysis, comprising: difference calculation means for calculating data.   2. The analysis skin image generation according to claim 1, wherein the difference image data is image data for analyzing skin spots, and the red component image data is image data for analyzing skin texture. system.   The difference calculating unit adjusts the amplitudes of waveforms that are approximate to each other in the red component image data and the blue component image data extracted by the extracting unit so as to be substantially the same, and then the difference image 3. The analysis skin image generation system according to claim 1, wherein data is obtained. The photographing means is composed of a camera mounted on a portable electronic device, and a skin photographing conversion lens provided in a lens module attached to the camera,
The skin image generation system for analysis according to any one of claims 1 to 3, wherein the illumination unit is provided in the lens module. A skin image generation method for analysis that generates skin image data capable of analyzing skin spots,
Extracting red component image data and blue component image data from image data taken by illuminating the subject with illumination light including at least red light and blue light,
After matching the brightness levels of the extracted red component image data and the blue component image data, difference image data that is the difference between the blue component image data and the red component image data is obtained. A method for generating a skin image for analysis, characterized in that it is obtained.   6. The analysis skin image generation according to claim 5, wherein the difference image data is image data for analyzing skin spots, and the red component image data is image data for analyzing skin texture. Method.   2. The amplitudes of waveforms that are approximate to each other in the extracted red component image data and the blue component image data are adjusted to be substantially the same before obtaining the difference image data. The analysis skin image generation method of Claim 5 or Claim 6.

Images