JP4909686B2 - Epidermal tissue quantification apparatus and program - Google Patents

Description

  The present invention relates to an epidermis tissue quantification apparatus and program for quantifying a subject's epidermis tissue.

  Conventionally, the skin condition has been quantitatively evaluated based on skin images.

  The skin image processing apparatus of Patent Document 1 evaluates wrinkles present on the surface of the skin using a three-dimensional model based on a digital image of the skin. The skin image processing apparatus plots the deepest part of the wrinkle based on the two face images, counts the edge width of the wrinkle, the distance between the deepest part and the edge in pixels, and determines the maximum based on the amount of movement. Calculate the depth.

  The skin surface condition evaluation apparatus of Patent Document 2 displays on the monitor screen an enlarged image at each of the standard site (ideal skin) and the evaluation site, the numerical value of the average area of the skin and the average depth of the skin groove, or a comparison result thereof. To do.

In the skin discrimination method of Patent Document 3, the state of the skin is taken into a color image, thinning processing is performed on a single color pixel image or a processed pixel image, and the peak width interval of the thin line is used as an index.
JP 2004-321793 A JP 2005-34424 A JP 2006-61170 A

  However, in Patent Documents 1 and 2, the two skins are only compared by requiring two images or monitoring the ideal skin in advance, and the epidermis tissue is not quantified. Moreover, the technique of patent document 3 only calculates an average value, a standard deviation, etc. from a brightness | luminance histogram, and does not quantify an epidermis tissue in detail.

  The present invention has been proposed to solve the above-described problems, and an object thereof is to provide an epidermis tissue quantification apparatus and program capable of accurately quantifying epidermis tissue.

  The epidermal tissue quantification device according to the present invention includes an imaging unit that images epidermal tissue, and a luminance value of each pixel of the image so that a variation in luminance value of the image captured by the imaging unit is equal to or greater than a predetermined value. A luminance value converting means for converting the luminance value, a binarizing means for binarizing the luminance value of each pixel converted by the luminance value converting means, a black pixel binarized by the binarizing means, and a predetermined value Short straight line matching means for matching the short straight lines and extracting the skin grooves represented by the connection of the matched short straight lines, and the number of short straight lines connected from the skin grooves extracted by the short straight line matching means is predetermined. The main skin groove extraction means for extracting the skin groove more than the value as the main skin groove, and the average thickness of the skin groove as an index for quantifying the epidermal tissue based on the skin groove extracted by the main skin groove extraction means , Skin gap distance, skin groove flat And a, and quantification index calculating means for calculating at least one of degree.

  Further, the epidermis tissue quantification program according to the present invention provides a computer with a luminance value for converting the luminance value of each pixel of the image so that the variation in the luminance value of the image captured by the imaging unit is equal to or greater than a predetermined value. Matching a conversion means, a binarization means for binarizing the luminance value of each pixel converted by the luminance value conversion means, and a black pixel binarized by the binarization means and a predetermined short straight line And a short straight line matching means for extracting a skin groove represented by the connection of the matched short straight lines, and a skin groove whose number of short straight lines is greater than or equal to a predetermined value from the skin grooves extracted by the short straight line matching means. The main skin groove extraction means for extracting as the main skin groove, and based on the skin groove extracted by the main skin groove extraction means, as an index quantifying the epidermal tissue, the average thickness of the skin groove, the interval of the skin groove Of the parallelism of the skin groove And quantification index calculation means for calculating one even without, thereby to function.

  The epidermal tissue quantification apparatus and program according to the present invention can accurately quantify epidermal tissue.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of an epidermis tissue quantification apparatus according to an embodiment of the present invention. The epidermis tissue quantification apparatus includes a camera 10 that images the human epidermis and a computer 20 that quantifies the epidermis based on the epidermis image captured by the camera 10. The computer 20 includes a CPU 21 that performs arithmetic processing, a RAM 22 that is a data work area, and a ROM 23 that stores a control program for the CPU 21. The epidermal tissue quantification apparatus configured as described above quantifies the epidermal tissue by performing the following processing.

  FIG. 2 is a flowchart showing an epidermis tissue quantification routine executed by the CPU 21 of the computer 20.

  In step S1, the CPU 21 of the computer 20 reads the skin image captured by the camera 10, and proceeds to step S2.

  In step S2, the CPU 21 executes a luminance conversion process for each pixel of the skin image. Here, the following luminance conversion routine is executed.

(Brightness conversion routine)
FIG. 3 is a flowchart showing a luminance conversion routine.

  In step S11, the CPU 21 calculates the average luminance value of the skin image, and proceeds to step S12. In the present embodiment, the luminance range of each pixel constituting the skin image is 0 to 255.

  In step S12, the CPU 21 generates a histogram representing the luminance distribution from the skin image, shifts the entire histogram so that the average luminance value becomes 128, which is the center value of the luminance range, and proceeds to step S13.

  In step S13, the CPU 21 reduces the histogram intercept width (W) representing the width of the histogram on both sides centered on the luminance average value (= 128) by a predetermined number (5 in this embodiment), and proceeds to step S14. Note that the initial value of the histogram intercept width is 127. Therefore, at first, the histogram intercept width is 127 minus 127, and as a result, the luminance range of the histogram is 6 to 249.

  In step S14, the CPU 21 calculates a standard deviation in the new luminance range of the histogram, and determines whether the standard deviation is 50 or more. If the determination is affirmative, the process proceeds to step S15. If the determination is negative, the process returns to step S13.

  In step S15, the CPU 21 performs linear conversion of the luminance value so that the luminance range of the histogram is from 0 to 255, and ends this routine. Thereby, step S2 shown in FIG. 2 is completed.

  4 to 6 are diagrams showing histograms of luminance values of the skin image. (A) Histogram intercept width W (new luminance range) and (B) standard deviation S.D. in the new luminance range. D. FIG. As shown in these figures, when the histogram intercept width W is 62 (luminance range 64 to 192), the standard deviation S.I. D. Becomes a value exceeding 51 (51). Then, the CPU 21 performs linear conversion of the luminance value so that the luminance range is from 64 to 192 to 0 to 255. As a result, even when the luminance values of the initial skin image are biased, the luminance values are appropriately varied, and high-precision image processing is possible. Then, the process proceeds to step S3 shown in FIG.

(Binarization processing)
In step S3, the CPU 21 executes binarization processing. In the present embodiment, in consideration of color unevenness, the binarization process is performed by changing the threshold value for each pixel to be binarized. Specifically, the following binarization processing routine is executed.

  FIG. 7 is a flowchart showing a binarization processing routine.

  In step S <b> 21, the CPU 21 matches the intersection of the cross area with the pixel (target pixel) to be binarized. Then, the CPU 21 calculates an average luminance value of each pixel in the cross area, and a luminance difference (= luminance average value−target pixel) that is a difference between the luminance value of the target pixel and the average luminance value of each pixel in the cross area. Brightness value), and the process proceeds to step S22.

  8 and 9 are diagrams showing a state in which the target pixel is set at the intersection of the cross areas.

  In step S22, the CPU 21 determines whether or not the luminance difference is equal to or greater than the binarization threshold. If the determination is affirmative, the process proceeds to step S23. If the determination is negative, the process proceeds to step S24.

  In step 23, the CPU 21 determines that the target pixel is a black pixel, and proceeds to step 25. In step 24, the CPU 21 determines that the target pixel is a white pixel, and proceeds to step 25.

  In step 25, the CPU 21 recognizes black pixels around the target pixel as the same group, and proceeds to step 26.

  In step 26, the CPU 21 determines whether or not the number of black pixels in the group is equal to or greater than the region threshold value in order to remove black pixels due to noise. If the determination is affirmative, the process proceeds to step 27. Proceed to 28.

  In step 27, the CPU 21 determines that the target group is a black pixel. In step 28, the CPU 21 determines that the target group is a white pixel. When step 27 or step S28 ends, this routine ends and the process proceeds to step S4 shown in FIG.

(Short straight line matching)
Next, in step S4, the CPU 21 performs short straight line matching processing. Specifically, the following short straight line matching routine is executed.

  FIG. 10 is a flowchart showing a short straight line matching processing routine.

  In step S31, the CPU 21 starts scanning from, for example, the upper left of the binarized skin image to search for a new black pixel, sets the searched black pixel as a matching start pixel, and proceeds to step S32.

  In step S32, as shown in FIG. 11, the CPU 21 sets one end of a predetermined short straight line (for example, width t = 1 pixel, basic length l = 20 pixels) on the matching start pixel. Then, the CPU 21 rotates a predetermined short straight line by a predetermined angle Δθ (for example, 5 degrees) with the matching start pixel as an axis, and determines whether or not 70% or more of the short straight lines are black pixels. . If the determination is affirmative, the process proceeds to step S33. If the determination is negative, the process returns to step S31. In this embodiment, a numerical value of “70”% is used, but it is needless to say that the present invention is not limited to this.

  In step S33, the CPU 21 calculates the thickness of the skin groove (black pixel portion) at the position of the matching start pixel, and proceeds to step S34. This step S33 will be described later in detail (see (Calculation of (average thickness) in step S5).

  In step S34, as shown in FIG. 12, the CPU 21 extends the length of the predetermined short straight line according to the thickness of the skin groove calculated in step S33. For example, the thicker the skin groove, the higher the extension rate of the short straight line. Then, the CPU 21 determines whether 70% or more of the extended straight line is a black pixel, and whether the pixel on the side opposite to the predetermined short straight line (the pixel at the end different from the matching start point) is an unprocessed pixel. If the determination is affirmative, the process proceeds to step S35. If the determination is negative, the process returns to step S31.

  In step S35, the CPU 21 matches the extended short straight line to the black pixel portion, and proceeds to step S36.

  In step S36, the CPU 21 sets the end pixel of the matched short straight line as a new matching start pixel, and proceeds to step S37.

  In step S37, the CPU 21 determines whether further matching processing can be performed with reference to the new matching start pixel. If the determination is affirmative, the process proceeds to step S38. If the determination is negative, the process returns to step S33.

  In step S38, the CPU 21 determines whether or not the matching start pixel has reached the search end point (for example, the lower right pixel when the binarized skin image is scanned from the upper left). When the determination is negative, the process returns to step S31. As a result of the execution of the short straight line matching routine, for example, nine short straight lines are connected as shown in FIG. Then, the process proceeds to step S5 shown in FIG.

(Main part extraction process)
In step S5, the CPU 21 extracts main skin grooves from the many skin grooves extracted in step S4. Specifically, the following main skin groove extraction routine is executed.

  FIG. 14 is a flowchart showing a main skin groove extraction routine.

  In step S41, the CPU 21 calculates the number of short straight lines in each group of skin grooves extracted in step S4, and proceeds to step S42.

  In step S42, the CPU 21 determines whether or not the number of short straight lines is equal to or greater than the connection threshold. If the determination is affirmative, the process proceeds to step S43. If the determination is negative, the process proceeds to step S44.

  In step S43, the CPU 21 determines that the skin groove of the group in which the number of short straight lines is equal to or greater than the connection threshold is the main skin groove. In step S44, the CPU 21 determines that the skin groove of the group whose short straight line connection number is not equal to or greater than the connection threshold is a non-main skin groove. And it progresses to step S45 through step S43 or step S44.

  In step S45, the CPU 21 extracts only the main skin groove and ends this routine. Then, the process proceeds to step S6 shown in FIG.

(Quantitative analysis)
In step S6, the CPU 21 performs a quantitative analysis of the epidermal tissue. Specifically, the CPU 21 calculates the average thickness, interval, and parallelism of the skin grooves.

(Average thickness)
The average thickness can be calculated in step S33 described above. First, the CPU 21 rotates the short straight line by 5 ° around the matching start pixel, and as shown in FIG. 15, the length from the center to the skin groove edge at each angle (number of continuous black pixels) t (θ ). Then, the sum of t (θ) and t (θ + 180) is T (θ), and the minimum value Tmin of T (θ) is the skin groove thickness T [pixel] at the matching start pixel as shown in FIG. . The CPU 21 repeats this every time the short straight line matching is performed, and as shown in FIG. 17, integrates the thicknesses at the respective matching start pixels (T _{1 to} T ₅ ). Finally, the CPU 21 calculates the average thickness Tl [pixel] of the skin groove in the image by dividing the integrated thickness by the number of short straight lines. Therefore, when the total number of matching short straight lines is N, the CPU 21 calculates the average thickness Tl of the skin groove according to the following equation (1).

（間隔ＬＩ）
一般に、皮溝の長さに注目すると、皮溝の全長が大きいほど、皮溝間隔が狭くなる傾向がある。ここで、皮溝の太さは間隔に対して微小であるので、無視することができる。よって、皮溝の平均太さは考慮せず、皮溝全長が皮溝間隔を表す指標となる。すなわち、ＣＰＵ２１は、次の式（２）に従って、皮溝間隔Ｌｌ［×１０^４pixel］を計算する。

(Interval LI)
In general, when attention is paid to the length of the skin groove, the gap length tends to become narrower as the overall length of the skin groove increases. Here, since the thickness of the skin groove is minute with respect to the interval, it can be ignored. Therefore, the average thickness of the skin groove is not taken into consideration, and the overall length of the skin groove is an index representing the skin groove interval. That is, the CPU 21 calculates the skin groove interval Ll [× 10 ⁴ pixels] according to the following equation (2).

なお、Ｓｌは、短直線マッチング画像の皮溝の面積であり、短直線マッチング画像におけるマッチング画素数とする。

Note that Sl is the area of the skin groove of the short straight line matching image, and is the number of matching pixels in the short straight line matching image.

(Parallelity, coefficient of variation)
The parallelism of the skin groove is used as an index representing the directionality of the skin groove.

18 and 19 are diagrams showing histograms representing (A) two skin images having different degrees of parallelism and (B) the number of short straight lines for each angle. In the skin image having a high degree of parallelism shown in FIG. 18A, the peak of the histogram is clear as shown in FIG. 18B, and the peak is generated at the angle indicating the direction of the skin groove. In contrast, in the image with low parallelism shown in FIG. 19A, the histogram has a gentle shape as shown in FIG. 19B, and the peak position is unclear. Therefore, the clarity of the peak of the histogram is represented by the standard deviation σ _n of the histogram height, which is used as an index of parallelism. Therefore, the CPU 21 calculates the standard deviation σ _n of the histogram height according to the following equation (3).

σ_ｎが大きいほど平行度が高いことを示す。なお、ｎ［θ］は角度θでの短直線本数、ＡＶＥ［ｎ［θ］］はｎ［θ］の平均値、Ｍはθの数（ヒストグラムの階級数）である。

It shows that parallelism is so high that (sigma) _n is large. Note that n [θ] is the number of short straight lines at the angle θ, AVE [n [θ]] is the average value of n [θ], and M is the number of θ (the number of classes in the histogram).

Here, even when the parallelism of the skin grooves is equal, σ _n may be greatly different if the number of short straight lines n [θ] is different. Therefore, in order to normalize the difference in the number of short straight lines n [θ], a value obtained by dividing the standard deviation σ _n by the average value of the number of short straight lines, that is, the coefficient of variation C _v may be used as the parallelism of the skin groove. In this case, CPU 21 may be calculated variation coefficient _{C v} according to the following equation (4).

なお、本発明は、上述した実施の形態に限定されるものではなく、特許請求の範囲に記載された範囲内で設計上の変更をされたものにも適用可能であるのは勿論である。

Note that the present invention is not limited to the above-described embodiment, and it is needless to say that the present invention can also be applied to a design modified within the scope of the claims.

It is a block diagram which shows the structure of the epidermis tissue quantification apparatus which concerns on embodiment of this invention. It is a flowchart which shows the epidermis tissue quantification routine which CPU of a computer performs. It is a flowchart which shows a brightness | luminance conversion routine. It is a figure which shows the histogram of the luminance value of an epidermis image, (A) Histogram intercept width W (new luminance range) and (B) standard deviation S.D in new luminance range. D. FIG. It is a figure which shows the histogram of the luminance value of an epidermis image, (A) Histogram intercept width W (new luminance range) and (B) standard deviation S.D in new luminance range. D. FIG. It is a figure which shows the histogram of the luminance value of an epidermis image, (A) Histogram intercept width W (new luminance range) and (B) standard deviation S.D in new luminance range. D. FIG. It is a flowchart which shows a binarization process routine. It is a figure which shows the state by which the object pixel was set to the intersection of a cross area. It is a figure which shows the state by which the object pixel was set to the intersection of a cross area. It is a flowchart which shows a short straight line matching process routine. It is a figure which shows the state which sets and rotates one end of a predetermined | prescribed short straight line on a matching start pixel. It is a figure which shows the state which extends the length of a predetermined | prescribed short straight line. It is a figure which shows the state in which nine short straight lines were connected. It is a flowchart which shows a main skin groove extraction routine. It is a figure which shows the state which measures the length (number of continuous black pixels) t ((theta)) from a center to a skin groove edge. It is a figure which shows the skin groove thickness T [pixel] in a matching start pixel. It is a diagram showing a state of accumulating the thickness of each matching start pixel _(T 1 _{~T 5).} It is a figure which shows the histogram which represented (A) skin image and (B) the number of the short straight lines for every angle. It is a figure which shows the histogram which represented (A) skin image and (B) the number of the short straight lines for every angle.

Explanation of symbols

10 Camera 20 Computer 21 CPU
22 RAM
23 ROM

Claims (4)

Imaging means for imaging epidermal tissue;
Luminance value conversion means for converting the luminance value of each pixel of the image so that the variation of the luminance value of the image captured by the imaging means is a predetermined value or more;
Binarization means for binarizing the luminance value of each pixel converted by the luminance value conversion means;
Short line matching means for matching a black pixel binarized by the binarization means and a predetermined short straight line, and extracting a skin groove represented by a connection of the matched short straight lines;
Main skin groove extracting means for extracting, as a main skin groove, a skin groove in which the number of short straight lines is not less than a predetermined value from the skin grooves extracted by the short straight line matching means;
Quantification for calculating at least one of the average thickness of the skin groove, the space between the skin grooves, and the parallelism of the skin groove as an index for quantifying the epidermal tissue based on the skin groove extracted by the main skin groove extracting means. Chemical index calculation means,
An epidermis tissue quantification apparatus comprising: The short straight line matching means sets the black pixel as a matching start pixel, sets one end of the short straight line as a matching start pixel, and rotates the short straight line by a predetermined angle around the matching start pixel as a predetermined ratio or more of the short straight line. The epidermal tissue quantification apparatus according to claim 1, wherein when the pixel becomes a black pixel, it is determined that the black pixel matches a predetermined short straight line. Computer
Luminance value conversion means for converting the luminance value of each pixel of the image so that the variation of the luminance value of the image captured by the imaging means is equal to or greater than a predetermined value;
Binarization means for binarizing the luminance value of each pixel converted by the luminance value conversion means;
Short line matching means for matching a black pixel binarized by the binarization means and a predetermined short straight line, and extracting a skin groove represented by a connection of the matched short straight lines;
Main skin groove extracting means for extracting, as a main skin groove, a skin groove in which the number of short straight lines is not less than a predetermined value from the skin grooves extracted by the short straight line matching means;
Quantification for calculating at least one of the average thickness of the skin groove, the space between the skin grooves, and the parallelism of the skin groove as an index for quantifying the epidermal tissue based on the skin groove extracted by the main skin groove extracting means. Chemical index calculation means,
Epidermis tissue quantification program to function. The short straight line matching means sets the black pixel as a matching start pixel, sets one end of the short straight line as a matching start pixel, and rotates the short straight line by a predetermined angle around the matching start pixel as a predetermined ratio or more of the short straight line. The epidermal tissue quantification program according to claim 3, wherein when the pixel becomes a black pixel, it is determined that the black pixel matches a predetermined short straight line.

Images