JP2001112722A - Surface state analyzing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a surface state analyzing method intended for a living body and capable of obtaining spectrograms at a number of wavelength points. SOLUTION: The constituent score of a main constituent is calculated, for each picture element, from data on a characteristic vector predetermined on the basis of data on spectral reflectance, as well as from spectrograms with dominant wavelengths of 400 nm, 550 nm and 700 nm, obtained by use of a trispectral camera. Spectrograms at a number of wavelength points are calculated from the constituent scores and the value of the characteristics vector.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a surface condition analysis method for analyzing the surface condition of an object, and more particularly, to a method for displaying a distribution of pigmentation in the skin in a quasi-stereoscopic manner in a plurality of wavelength bands. The present invention relates to a surface state analysis method suitable for efficiently acquiring a spectral image.

[0002]

2. Description of the Related Art Since light on the long wavelength side can reach deeper into the skin than light on the short wavelength side in the visible region, a spectral image of the skin surface obtained in a specific wavelength band is applied to that wavelength. It has been shown that it represents the distribution of melanin at the corresponding depth (Osamu Kaneko et al., Cosmetic Journal, 32
(4) , pp 361-371, 1998; Yukiko Kawaguchi et al., Cosmetic Journal, 33 (1) , pp 27-38, 199.
9).

In the above-mentioned paper, the main wavelength is 400 nm,
Spectral images at 550 nm and 700 nm have been used to represent the distribution of melanin at depths corresponding to each wavelength. However, if spectral images of more wavelengths are obtained, it is expected that more useful information such as the three-dimensional distribution of melanin inside the skin will be obtained.

Recently, ten interference filters have been replaced with C
A system has been reported in which a spectral image at 10 wavelength points can be obtained by sequentially rotating at high speed in front of a CD (Yasuhiro Komiya et al., Image Lab., 7 (5) , pp. 37-40, 199 ) .
6). This system is a color classification camera system developed to classify subtle color differences in images. Furthermore, 3
Spectral data of 50 × 350 points (400 nm to 100
A two-dimensional image spectroscopy system capable of obtaining a single measurement at 0 nm and 10 nm intervals has also appeared on the market ( Tadao Miura et al., Imaging Lab., 8 (11) , pp. 54-58, 1997). In this method, a spectrum is calculated by performing an inverse Fourier transform on an interferogram obtained for each pixel. However, it requires several seconds to several minutes of measurement time, and is not suitable for collecting a spectral image of a living body. This device was developed mainly for detecting abnormalities in the structure and number of chromosomes in genetic diagnosis, cytohistology, and the like.

[0005]

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a surface state analysis method suitable for collecting a spectral image of a living body and capable of obtaining a spectral image at a number of wavelength points. is there.

[0006]

According to the present invention, spectral images in the first plurality of wavelength bands are acquired for the same object, and the main component of the spectrum of light reflected from the object is obtained. The component score is determined from the values of the pixels constituting the spectral image in the first plurality of wavelength bands, and the determined second component wavelengths are used to determine the second plurality of wavelengths greater than the first plurality of wavelength bands. There is provided a surface state analysis method including a step of calculating each of spectral images in a band.

[0007] Preferably, the method further comprises the step of generating a three-dimensional display using wavelengths as third coordinate axes from the spectral images in the second plurality of wavelength bands.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION First, assuming a white female skin as an example of weak coloration (light color) due to melanin and a black female skin as an example of strong coloration (dark color), ,
From the spectral reflectance data of the cheeks of black and yellow females, the number of spectral components and their magnitudes are determined assuming a linear structure.

[0009] destination reported by Case (Yukiko Kawaguchi other,粧技Journal, 33 (1), pp27-38,1999) number of constituent factors V _i of the resulting skin spectral reflectance at it is 3.
Therefore, the spectral reflectance R (λ) at an arbitrary wavelength is given by the equation (1)
Can be written in the form of _{R (λ) = M 1 V} 1 (λ) + M 2 V 2 (λ) + M 3 V 3 (λ) (1) Here, M _i: weighting factor related to the structure factor V _i V _i (λ): Wavelength the size of the structure factor V _i in lambda Incidentally, the problem of extracting the structure factor of the spectral reflectance R (lambda) in the form of equation (1) is between the wavelengths of the spectral reflectance of the skin, raw sum of squares, the product It is attributable to the problem of eigenvalues and eigenvectors of the union matrix. Thus the weighting coefficient M _i configuration factor V _i in equation (1), the magnitude V _i (lambda) is the structure factor V _i in is the wavelength lambda component scores obtained in the principal component analysis, the eigenvector itself.

[0010] The skin is photographed with a trispectral camera described in Japanese Patent Application Laid-Open No.
Obtain spectral images at m, 550 nm and 700 nm. Next, the method reported earlier (Yukiko Kawaguchi et al., Cosmetic Magazine, 33 (1) , p.
(pp. 27-38, 1999), the luminance of each pixel is converted into a spectral reflectance. Assuming that the obtained spectral reflectances for each pixel are R ₄₀₀ , R ₅₅₀ and R ₇₀₀ , equation (1) becomes equation (2)
Can be written as

R ₄₀₀ = M ₁ V _1,400 + M ₂ V _2,400 + M ₃ V _3,400 R ₅₅₀ = M ₁ V _1,550 + M ₂ V _2,550 + M ₃ V _3,550 (2) R ₇₀₀ = M ₁ V _1,700 + M ₂ V _2,700 + M ₃ V _3,700 Solve the simultaneous equations of equation (2), calculate M ₁ , M ₂ , and M ₃ for each pixel, substitute them into equation (1), and apply them to the visible region (380 nm to 780 nm, 10 nm interval, 41 wavelengths). The spectral reflectance R (λ) for each pixel at the point (dot) is determined.

The obtained maximum value R _{max of the} spectral reflectance for each wavelength.
(Λ), the minimum value R _min (λ), and the spectral reflectance R (λ) of the pixel at the wavelength point are substituted into Equation (3) to obtain the spectral reflectance R (λ) of the pixel at the wavelength. Brightness I (λ)
(0-255). I (λ) = 255 × (R (λ) −R _min (λ) / (R _max (λ) −R _min (λ)) (3) Spectral image using I (λ) thus obtained creating a. the value of V _i (λ) is obtained for 41 wavelength points Using data of the spectral reflectance of 41 wavelength points in the calculation of V _i (λ), the spectral of 41 wavelength points with it An image can be obtained, but if more spectral points or spectral images of arbitrary wavelengths are required, the tertiary natural spline interpolation is performed using the spectral reflectance data of the above-mentioned 41 wavelength points corresponding to individual pixels. The spectral reflectance R (λ) of the pixel at the wavelength point is calculated based on the equation, and is substituted into the equation (3) to obtain the luminance I.
Based on (λ), a spectral image is created.

It is preferable to remove pores and noises in the obtained spectral image by image processing, and to fill the holes using the average luminance of the surroundings. The binarization of the spectral image created by the method described above is performed according to the method of Otsu (Nobuyuki Otsu, IEICE Transactions , J63-D (4) , pp349-356, 1 ) .
980), but it is desirable to correct (−20) the calculated threshold since the size of pigmentation tends to be extracted slightly larger.

[0014] By stacking and displaying the obtained 41 binary images in order of wavelength, a cluster of melanin existing on the surface layer of the skin is visualized in a pseudo manner. A 380 nm binary image is displayed with the top surface of the cube shown in FIG. 1 corresponding to the skin surface. With the −Z direction as the depth direction, the obtained binary images are stacked in order of wavelength, and the lowermost surface is a 780 nm binary image. Since a three-dimensional effect cannot be obtained simply by displaying a binary image,
Perform shading.

Further, an image obtained by cutting out the shape of melanin existing in the surface layer of the skin at an arbitrary cross section in the X, Y, and -Z axial directions may be displayed on a screen.

[0016]

EXAMPLE Assuming that the skin of a white woman is assumed as an example of weak coloration (light color) due to melanin and the skin of a black woman as an example of strong coloration (dark color), a Hitachi Color Analyzer 607 is used.
In total, there are 75 women (whites, 12 blacks, 63 yellows (Japanese and Southeast Asians);
The spectral reflectance of the cheek was measured. Previous report (Yukiko Kawaguchi et al., Cosmetic Magazine, 32 (1) , pp. 27-38, 199 )
In 9), the obtained 380 nm to 780 nm, 10 nm interval, 4
The spectral reflectance values at 31 wavelength points between 400 nm and 700 nm are extracted from the spectral reflectance data at one wavelength point, and the principal component analysis between wavelengths and the raw sum of squares and sum of products matrix is performed to extract the reflected light constituent factors. are doing.

In the present specification, in order to obtain a spectral image at a shallower place and a deeper place, measured 380 nm to 780 nm,
The above-described principal component analysis was performed on spectral reflectance data at 41 nm points at intervals of 10 nm. Table 1 shows the obtained eigenvalues and cumulative contribution rates.

[0018]

[Table 1]

As can be seen from Table 1, equation (1) also holds in this case. FIG. 2 shows the obtained eigenvectors. Table 2 shows the maximum value and the minimum value of the reflectance for each wavelength for converting the reflectance R (λ) of the pixel at the wavelength point into the luminance I (λ) using Expression (3).

[0020]

[Table 2]

The main wavelength 400 of freckles, liver spots, bears, and moles obtained using the above-mentioned trispectral camera.
Based on the spectral images at nm, 550 nm, and 700 nm, spectral images at 380 nm to 780 nm, at 10 nm intervals, and at 41 wavelength points were reconstructed by the method of the present invention. 3 reconstructed in FIGS.
8 shows spectral images of 80 nm to 780 nm (10 nm intervals, 41 wavelength points).

FIGS. 7 to 10 show the reconstructed freckles, liver spots, bears, and moles, which are binarized into spectral images at 41 wavelength points, and the binary images are sequentially laminated and shaded.
Shown in As described above, the binarized image on the shorter wavelength side is arranged above each figure, and shows the distribution of melanin existing in a relatively shallow place. Further below, a binary image on the longer wavelength side is arranged, and the distribution of melanin present at a relatively deep place is displayed. The presence of melanin forming freckles can be confirmed at relatively shallow places, and the presence of melanin forming melasma and dark spots can also be confirmed at deeper places. In the case of a mole, a series of melanin clusters can be seen from a shallow place to a deep place.

FIGS. 11 to 13 show the results of visualizing the melanin on the skin surface caused by transient sunburn by the method of the present invention. FIG. 11 shows the results of visualization of melanin on tanned skin whose redness has disappeared one week after sunburn,
Figures 12 and 13 show the results after 2 weeks and 4 weeks, respectively. As can be seen from FIGS. 11 to 13, the process of the present invention makes it possible to visualize the process in which melanin migrates to the upper layer over time and the melanin generated by transient sunburn disappears.

[0024]

As described above, according to the present invention, it is possible to obtain a spectral image of a living body at many wavelength points.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a three-dimensional display generated from spectral images in many wavelength bands.

FIG. 2 is a diagram showing eigenvectors obtained in advance from spectral reflectance data.

FIG. 3 is a diagram showing an example of a spectral image reconstructed by calculation according to the method of the present invention.

FIG. 4 is a diagram showing an example of a spectral image reconstructed by calculation according to the method of the present invention.

FIG. 5 is a diagram showing an example of a spectral image reconstructed by calculation according to the method of the present invention.

FIG. 6 is a diagram showing an example of a spectral image reconstructed by calculation according to the method of the present invention.

FIG. 7 is a diagram showing an example of visualizing the distribution of melanin present in the skin surface layer according to the method of the present invention.

FIG. 8 is a view showing an example in which the distribution of melanin present in the skin surface layer is visualized according to the method of the present invention.

FIG. 9 is a diagram showing an example in which the distribution of melanin present in the skin surface layer is visualized according to the method of the present invention.

FIG. 10 is a diagram showing an example in which the distribution of melanin present in the skin surface layer is visualized according to the method of the present invention.

FIG. 11 is a diagram showing an example of visualizing the distribution of melanin present in the skin surface layer according to the method of the present invention.

FIG. 12 is a diagram showing an example of visualizing the distribution of melanin present in the skin surface layer according to the method of the present invention.

FIG. 13 is a diagram showing an example of visualizing the distribution of melanin present in the skin surface layer according to the method of the present invention.

Claims (2)

[Claims] 1. A spectral image in a first plurality of wavelength bands is acquired for the same target object, and component scores of main components included in a spectrum of light reflected from the target object constitute the spectral image. A surface state analysis method comprising the steps of: determining a spectral image in a second plurality of wavelength bands greater than the first plurality of wavelength bands by using the determined component score, and determining the spectral images by using the determined component scores. . 2. The surface state analysis method according to claim 1, further comprising the step of generating a three-dimensional display using wavelengths as third coordinate axes from the spectral images in the second plurality of wavelength bands.