JP6266948B2 - Jaundice measurement system - Google Patents

Description

The present invention relates to a yellow疸measurement system.

  The measurement of jaundice is one of the most common blood tests performed during the neonatal period, and at the same time is an emphasis that medical personnel and parents closely observe in the early days of the baby. If the jaundice value is too high, the newborn may develop nuclear kernicterus and cause brain damage, and in severe cases, there may be sequelae such as cerebral palsy, hearing loss and intellectual disability.

  Currently, jaundice measurement methods depend mainly on blood sampling or testing with a small amount of blood from the footpad, but this step belongs to an invasive test, so there are side effects such as pain, bleeding and infection from wounds. There was a risk of it happening, and there were restrictions on implementation and anxiety to the family. Therefore, most blood tests have been confirmed by blood tests after discovering that jaundice has appeared on the baby's skin through macroscopic observation of family members or medical personnel. However, macroscopic observation of jaundice is a subjective judgment, especially since the parents of newborns have not received professional training and have no experience. I have an excessive blood test. In addition, jaundice meters can irradiate the baby's skin with the beam and analyze the wavelength of the returned light to provide non-invasive jaundice readings. This instrument is not widely applied in the world, as it cannot be borne by a typical family and most of the time requiring jaundice measurement is only within one to two months after the baby was born.

JP 2012-61232 A

An object of the present invention is to provide an objective jaundice index while performing measurement on a newborn's skin by an image processing (non-invasive) method through a smartphone, a tablet computer or a personal computer with a camera function which is very popular at present. in, it is possible to occasionally self-monitoring indicators neonatal jaundice parents of newborn, when an indication of jaundice is higher than the normal value, to provide a yellow疸measurement system capable Hassu return-visit warning message is there.

A jaundice measurement system according to the present invention includes an electronic device in which a program capable of executing the following jaundice measurement method is installed. The jaundice measurement method includes step S1, step S2, and step S3 . Step S1 captures images of the newborn's face and color chart. In step S2, color correction is performed on the captured image in accordance with the color chart. In step S3, the jaundice index is calculated based on the color information of the face pixels of the image and the defective image is excluded.
In step S1, an image is acquired by an electronic device equipped with a camera and an executable program.
Step S2 includes step S21 for automatically detecting the position of the color chart, and step S22 for performing color normalization on the entire image based on the color average value of the color chart.
Step S3 includes step S31, step S32, and step S33.
A step S31 searches for a pixel belonging to the skin at a recommended face position based on a normal skin color hue range. In step S32, pixels belonging to the jaundice region are detected based on a predetermined hue range in the skin pixels. Step S33 calculates the ratio between the number of pixels in the jaundice area and the number of pixels in the skin area. In step S34, the ratio between the number of pixels in the jaundice area calculated in step S33 and the number of pixels in the skin area is converted into a jaundice index.
In the photographing process, it is determined whether or not the newborn's complexion is abnormal based on the normal range of jaundice at the time of birth of the newborn.

  The step S1 for capturing the image of the newborn's face and color chart is photographed with a tool such as a smartphone, a tablet computer or a personal computer, and capable of executing an application program (APP) of the method according to the present invention. The vote has three colors, pure red, pure green, and pure blue. To determine the composition, set the face outline and recommended color chart outline on the preview screen of the shooting, and load the newborn's postnatal time in the shooting process before or after shooting. This corresponds to the determination of the normal range of jaundice. The color chart is rectangular and can be arranged around the position of the newborn's face, so that the background color of the newborn's face, such as swaddles, duvets or pillows, is red, green, blue, so as not to interfere with image analysis. Avoid colors such as yellow.

Step S2 for calibrating the color of the captured image in accordance with the color chart is to use the RGB color model and perform color normalization on the entire image based on the average value of the color chart color. In step S2, the position of the color chart is automatically searched to calculate the total number N _x of pixels of the color chart area, x∈ {r, g, b} is set, and the N _x exceeds the set value. In addition, step S21 for calculating the color average value of the color chart and step S22 for performing color normalization on the entire image based on the color average value of the color chart are further included. In step S21, after the image is captured, the color regions that are very close to pure red, pure green, and pure blue are searched within the outline of the color chart at the recommended position, and the maximum connecting range of these regions is found to detect red, The green and blue color chart areas are used.

The average value of the red channel (Red Channel) in the in the region where there is a color chart red after searching in the present invention is defined as a color P _r in captured images red color chips. Color processing method of the green and blue of the captured image of the color chart be the same, it is possible to obtain the P _g and P _b.

  In step S22, color normalization is performed on the entire image based on the color average value of the color chart, and each pixel in the image is adjusted for the red, green, and blue channels, and the color difference of the image that occurs in different shooting environments. Can be reduced.

  The step S3 of calculating the jaundice index based on the color information of the face pixels of the image automatically acquires a region belonging to the skin color of the face in the HSV (Hue, Saturation, Value) color model, and the region belonging to the jaundice in this region Is to search for. Calculate the average value of saturation of the pixels in the jaundice area and use it as a jaundice index corresponding to the jaundice value obtained by blood test through a special function transformation. Also, the number of pixels in the jaundice area and the number of pixels in the skin area This is the basis for image quality evaluation incorporating the ratio.

Step S3 is a step S31 of calculating the total number N _SK with searching pixels that belong to the skin in the face position recommended based on the hue (Hue) the scope of the general skin color hue range of further jaundice values in pixels of the skin A step S32 for finding pixels belonging to the jaundice area based on the above, a step S33 for calculating a ratio of the average value of saturation of these jaundice area pixels and the number of jaundice area pixels and the number of skin area pixels (Joundice SKin Ratio, JSR); Step S34 for converting the degree average value into the jaundice index.

In step S31, the hue range of the skin area of the newborn's face statistically determined by the inventor is set as the designated facial pixel parameter, and this image is captured within the captured oval face outline range. Are selected and the total number is _NSK . Further, in order to improve the quality and accuracy of the captured image, after step S31, the total number of pixels N _E in the contour line and the average saturation value S _E thereof are calculated and searched, and the ratio of N _SK to N _E (SKin Ellipse Ratio) , SER) can proceed to step S311. Furthermore it is possible to proceed to step S32 when the S _E and SER value exceeds the set value.

  In step S32, the hue range of the neonatal jaundice area calculated by the inventor is used as a set jaundice pixel parameter to further select the jaundice area pixel of the image in the captured face range pixels.

  In step S33, a ratio between the total number of jaundice area pixels and the total number of facial skin area pixels is obtained. In order to avoid misjudgment due to poor image quality, it is recommended that either the total number of jaundice area pixels or the ratio of the two or the average saturation value of the pixels in the face outline range fall below the set value to recapture the image. .

  The step S34 calculates an average value of color saturation of the pixels in the jaundice region, converts it into a jaundice index, and correlates the obtained jaundice index with the jaundice value obtained by the blood test. When the index is higher than general medium risk, warn and recommend a revisit.

  The method of the present invention may be a computer program (APP) product. The computer program (APP) can be downloaded from the Internet or loaded into a smart phone with a camera function, a tablet computer, a personal computer, or an electronic device similar to these, which is currently very popular through the other route, thereby implementing the method of the present invention. Since it can be photographed in accordance with the color chart at the time of operation of the present invention and can be analyzed in real time online, it can have the effect of convenience of operation. In addition, the present invention requires more accurate measurement results because the quality of the captured image is required to be re-photographed when the requirements (for example, the light is too dark or the face and color chart positions are not ideal) are required. And the measurement conversion index of the imaging method of the present invention corresponds to the jaundice value of the blood test, which is more accurate than the visual measurement method, and does not miss the treatment timing due to a gross misjudgment to the newborn. Therefore, it is possible to alleviate the concerns of newborn parents, reduce excessive blood tests, avoid the inconvenience that the family could not find because the jaundice value was too high, and also reduce medical resources. is there.

It is a schematic diagram which shows the processing step of the jaundice measurement program by one Embodiment of this invention. It is a flowchart of the jaundice measurement program by one Embodiment of this invention. It is a schematic diagram which shows the process by the jaundice measurement program by one Embodiment of this invention. It is a characteristic view which shows the processing result by the jaundice measurement program by one Embodiment of this invention. It is a jaundice curve after birth of a newborn published by the American Academy of Pediatrics.

(One embodiment)
This will be described with reference to FIGS.
FIG. 1 is a diagram showing processing steps of a jaundice measurement program according to an embodiment of the present invention, FIG. 2 is a flowchart, and FIG. 3 is a diagram showing a state in which a photographed image of this embodiment is combined with a color chart. In the present embodiment, S1 that captures an image of a newborn's face and color chart, S2 that performs color calibration on the captured image in accordance with the color chart, and calculates a jaundice index based on color information of face pixels of the image. S3 for eliminating defective images.

The step S1 for capturing the image of the newborn's face and color chart is photographed with a camera such as a smartphone, a tablet computer or a personal computer or an electric device having a function similar to this and a tool capable of executing a program. The color charts are color checker-type color charts 11, 12, and 13 having three colors of pure red, pure green, and pure blue, and the dimensions of the color charts 11, 12, and 13 are all over 1 cm ² . In order to determine the composition shown in FIG. 3, the oval face outline 2 and the color chart outline 3 at the recommended positions of both rectangular color charts are set on the preview screen of the photographing. The green and pure blue color charts 11, 12, and 13 are arranged in the upper color chart outline 3, and the newborn's postnatal time data is loaded in the photographing process before or after photographing. 5. Corresponding to the determination of the current normal range of jaundice for newborns shown in FIG. Avoid background colors such as red, green, blue, yellow, etc. for objects around the newborn's face, such as swaddles, futons or pillows, so as not to interfere with image analysis.

Step S2 for calibrating the color of the captured image in accordance with the color chart is to use the RGB color model and perform color normalization on the entire image based on the average value of the color chart color. In step S2, the position of the color chart is automatically retrieved to calculate the total number N _x of pixels in the color chart area, and x∈ {r, g, b} is set, and the Nx exceeds the set value. The method further includes a step S21 for calculating the color average value of the color chart and a step S22 for performing color normalization on the entire image based on the color average value of the color chart. In step S21, after the image is taken in, the color chart contour line 3 at the recommended position of both rectangular color charts is searched for each of the color areas very close to pure red, pure green, and pure blue, and the maximum connection range of these areas is further determined. The red, green, and blue color charts 11, 12, and 13 are respectively found out.

The automatic retrieval of the color chart position of the present embodiment is to make mxRr _i , mxRg _i , and mxRb _{i be} the maximum continuous regions in Rr _i , Rg _i , and Rb _i , respectively, and max _i {mxRr _i , mxRg _i. , MxRb _i } includes the maximum mxRr _i , mxRg _i , and mxRb _i at the same time in the i-th region, and these regions are regions having pure red, pure green, and pure blue color charts 11, 12, and 13. In the present embodiment, the positions of the color charts 11, 12, and 13 can be searched for in the color chart outline 3 located above the face. The MxRr _i, in mxRg _i, mxRb _i of the number of pixels individual regions _{N x, x∈ {r, g} , b} and, when the N _x is below the set value th _Nx, color chart exactly color chart not placed in the contour 3, or shows a lack of image capture time of light, and the N _x value corresponding to the setting of the camera resolution, N _x value is increased at high resolution. For example, when the resolution is 4752 × 3168 and the th _Nx value is 10000 and is lower than 10,000, it is recommended that the color chart position is adjusted and then an image is captured again.

Also, {CP _i } is set as a set of pixels in the color chart outline 3 at the recommended position of the rectangular color chart, i∈ {0, 1}, and i does not belong to the fixed area. Rr _i, Rg _i, Rb _i each i th pure red within the rectangle, pure green, is pure blue and very close color region.

  In this embodiment, the default threshold value of the color chart color is set to th_cp. th_cp satisfies the following formula 1. Set the preferred th_cp to 10.

（数式１において、ｘ∈｛ｒ，ｇ，ｂ｝、（ｒ_r、ｒ_g、ｒ_b）＝（２５５、０、０）、（ｇ_r、ｇ_g、ｇ_b）＝（０、２５５、０）、（ｂ_r、ｂ_g、ｂ_b）＝（０、０、２５５）である。）
ａとｂ間の距離は下記の数式２で算出する。

(In Formula 1, x∈ {r, g, b}, (r r, r g, r b) = (255,0,0), (g r, g g, g b) = (0,255, 0), (b _r , b _g , b _b ) = (0, 0, 255).)
The distance between a and b is calculated by the following formula 2.

The average value of the red channel (Red Channel) in the in the region where there is a red color chip 11 after searching in this embodiment is defined as a color P _r in captured images red color chip 11. Color processing method of the green and blue of the captured image of the color chart be the same, it is possible to obtain the P _g and P _b.

  In the present embodiment, the two rectangular color chart outlines 3 are set as color charts that can be easily arranged by the user, but in the present embodiment, only one color chart outline 3 can be set. When capturing an image, if the color chart cannot be placed in the color chart outline 3 due to some force majeure, first place the area other than the face and other areas within the image capture range, and then manually capture the image. The procedure for setting the color chart position can be executed. That is, when the user manually clicks the center point of the color chart range, the present embodiment can automatically search for the color chart near the position. The color chart searching method is the same as the above-described step.

Next, step S22 for performing color normalization on the entire image based on the color average value of the color chart adjusts each pixel in the image with respect to the red, green and blue channels, respectively, and an area where the red color chart 11 is present. The red channel value of is close to 255. The processing method of the green and blue color charts 12 and 13 is the same. _{f r = 255 / P r,} f g = 255 / P g, if the f _b = 255 / P _b, adjusts the red channel C _r of each pixel in the image to _C'r = f _r × C _r The processing method of the green channel C _g and the blue channel C _b of each pixel is the same, and C ′ _g = f _g × C _g , C ′ _b = f _b × C _b, and C ′ _r , C ′ _g , C ′ _b RGB color information is acquired, the pixel values of each pixel in the image are calculated by the corresponding color charts 11, 12, and 13, and the color difference of the image due to different shooting environments can be reduced.

A description will be given with reference to FIG. 4 of a regression statistical chart corresponding to the bilirubin value of the average value of color saturation of the present embodiment. The step S3 of calculating the jaundice index based on the color information of the face pixels of the image automatically acquires a region belonging to the skin color of the face in the HSV (Hue, Saturation, Value) color model, and the region belonging to the jaundice in this region Is to search for. Calculate the average value of saturation of the pixels in the jaundice area and use it as a jaundice index corresponding to the jaundice value obtained by blood test through a special function transformation. Also, the number of pixels in the jaundice area and the number of pixels in the skin area (Jandice Skin Ratio, JSR) is used as the basis for evaluation of image quality. When the JSR value is lower than th _JSR , the quality of the captured image is poor and a relatively large error occurs in the calculated jaundice index. th _JSR is a default threshold (Threshold) obtained after the statistics of the present embodiment.

Step S3 is generally skin color hue and step S31 of calculating the total number N _SK with looking for pixel (pixel) belonging to the skin recommended face position based on (Hue) range, further jaundice value in the pixel of skin Step S32 for finding pixels belonging to the jaundice area based on the hue range of the image, and calculating the ratio of the saturation (Saturation) average value of these jaundice area pixels to the number of jaundice area pixels and the number of skin area pixels (Jandice SKin Ratio, JSR) Step S33, and step S34 for converting the saturation average value into the jaundice index.

The C ′ _r , C ′ _g , and C ′ _b (RGB) are converted into (HSV) by the following Equation 3.

  Since the conversion formula is a conventional technique, description thereof is omitted. The step S31 sets the hue range (Hue) range (about 42.5 ± 18.5 in this embodiment) of the skin area of the newborn's face as calculated by the inventor after conversion to the HSV color information by the RGB color information. By using the acquired face pixel parameters, the skin area pixel set SK of the face of this image is selected within the range 2 of the captured outline of the oval face.

  In step S31, if the face skin area can be within the range of the oval face contour 2 due to the force of force when capturing an image, the image is manually captured after capturing the image in this embodiment. By adjusting the position, size, and angle of the contour line 2 of the oval face, the skin area of the face of the captured image is covered within the range of the contour line 2 of the oval face.

In order to further improve the quality and accuracy of the captured image, the present embodiment calculates and retrieves the total number N _E of pixels within the contour line and the saturation average value S _E after step S31, and also calculates the ratio SER of N _SK and N _E. Step S311 may be further included. The ratio SER = N _SK / N _E of the total number N _SK of pixel sets SK of the skin region of the face and the total number N _E of pixel sets E within the range 2 of the oval face outline, and the SER value is the th _SER value. If the value is below, it is determined that the quality of the captured image is poor and it is recommended that the image be captured again. In order to automatically recognize whether or not the difference between the skin area of the face and the position 2 of the oval face outline is too large when the image is captured, the present embodiment captures when the SER value falls below th _SER (70%). However, it is recommended that the quality of the image is recognized as poor and that the image is read again. Further, the saturation average value of all the pixels within the range 2 of the oval face contour line is S _E , and in this embodiment, when the S _E value falls below th _SE (0.5), the image is captured. The setting is recommended to recognize that the ambient light is insufficient and to re-import the image. In the step S32, the hue range (about 40 to 60 in this embodiment) of the neonatal jaundice region calculated by the inventor is set as the pixel parameter of the jaundice, and the image of the image in the captured face range is further reduced. The jaundice region pixel set J is selected.

The step S33 obtains the ratio JSR = N _J / N _SK of the total number N _SK of the total number of jaundice region pixel set J N _J and face skin region pixel set SK. To avoid erroneous determination due poor quality of the image, this embodiment is recognized and captured image quality defect when JSR value is below the th _JSR (50%), a relatively large error in the calculated later J _S For this reason, it is recommended that the image be read again.

In step S34, a color saturation (Saturation) average value J _S of the pixel set J is calculated, converted into a jaundice index by a special function T (x), and the obtained jaundice index is converted into blood. It is equivalent to jaundice value J _V obtained in the inspection. As shown in FIG. 4, the function T (x) is a regression statistical functions J _S and J _V, and the J and _S and J _V is a proportional relationship, J _S also increases higher J _V. Referring with jaundice curve diagram the American Academy of Pediatrics corresponds to old times published by neonatal 5, it is determined that whether jaundice index higher too at postnatal age of the J _V values and newborn (postnatal age), The jaundice index warns and recommends a revisit when it exceeds the general moderate risk recommended by the American Academy of Pediatrics (eg, above the median curve) or greater than or equal to 15 mg / dL.

  This embodiment may be a computer program (APP) product. When the computer program (APP) is loaded from the Internet to a smartphone with a camera function, a tablet computer, a personal computer, or an electronic device similar to these, which is currently very popular, the method of this embodiment can be implemented.

  Since it can be photographed in accordance with the color chart during operation of this embodiment and can be analyzed in real time, it is not necessary to use light in the range of a special spectrum, or it is not necessary to photograph an image under a special environment. Can have an effect. In addition, providing imaging measurements to newborn parents at any time can reduce that concern and, at the same time, avoid the inconvenience of frequent blood tests in the hospital.

  The present invention requires a more accurate measurement result because the quality of the captured image is required to be retaken if the light does not meet the requirements (for example, the light is too dark or the face and color chart positions are not ideal). When the measurement conversion index of the imaging method of the present invention corresponds to the blood test jaundice value, is more accurate than the newborn parents' eye measurement method, and the jaundice index is higher than the normal value, a revisit warning message is issued. So newborns will not miss the timing of treatment due to gross judgment mistakes by newborn parents. The above-mentioned embodiment is to clarify the technical contents of the present invention, and is not limited by this example. It goes without saying that various improvements and modifications can be made within the scope of the present invention.

11 ... Red color chart,
12 ... Green color chart,
13 ... Blue color chart,
2 ... Face outline,
3 ... Color chart outline.

Claims (9)

Step S1 for capturing an image of a newborn face and color chart,
Step S2 for calibrating the color of the captured image according to the color chart, and
A method for measuring jaundice including step S3 for calculating a jaundice index based on color information of face pixels of an image and eliminating a defective image,
In step S1, an image is acquired by an electronic device equipped with a camera and an executable program,
Step S2 includes step S21 for automatically detecting the position of the color chart, and step S22 for performing color normalization on the entire image based on the color average value of the color chart.
The step S3 includes a step S31 of searching for a pixel belonging to the skin at a face position recommended based on a normal skin color hue range, and a step of finding a jaundice region in the skin pixel based on a predetermined hue range. Step S32 for detecting a pixel to which it belongs, Step S33 for calculating the ratio between the number of pixels in the jaundice area and the number of pixels in the skin area, and the ratio between the number of pixels in the jaundice area and the number of pixels in the skin area calculated in step S33 Converting S into a jaundice index, and
In imaging process, based on the jaundice normal range, jaundice measurement system comprising electronics a yellow疸measured way to determine whether complexion neonatal is abnormal executable program is installed. The color chart is a color checker that includes pure red, pure green, and pure blue.
The step S21 of automatically detecting the position of said color chart calculates the pixel total number N _x of the color chart area, when the total number of pixels N _x exceeds a predetermined value, to again calculate the color average value of the color chart Step,
The step S31 is a step of calculating the total number of skin pixels based on the range of the general skin color hue (Hue),
In step S311 executed between step S31 and step S32, if the quality of the image is poor, it is recommended that the image be re-captured.
In step S33, when the saturation average value of the pixels belonging to the jaundice area and the ratio (JSR) between the number of pixels in the jaundice area and the number of pixels in the skin area are calculated, and the value of the ratio falls below a predetermined value The jaundice measuring system according to claim 1, wherein the image is recaptured again. 3. The jaundice measuring system according to claim 2, wherein objects around the face of the newborn avoid red, green, blue and yellow colors. Set the face outline and at least one color chart outline on the shooting preview screen,
In step S21, the position of the color chart is automatically searched and the color chart is set on the outline of the color chart, i∈ {0, 1},
mxRr _i, searches the color chart to mxRg _i, mxRb _i is each Rr _i, Rg _i, the largest contiguous area in Rb _i,
max _i {mxRr _i , mxRg _i , mxRb _i } has the maximum mxRr _i , mxRg _i , and mxRb _i at the same time in the i-th region, and the color chart is located in pure red, pure green, and pure blue colors The area where the vote is located,
{CP _i } is a pixel set in the color chart outline,
Rr _i, and Rg _i, net red in the color chart contour Rb _i, pure green, pure blue and very close color region,
th_cp is a default threshold value of the color of the color chart and satisfies the following formula 1;

(In Formula 1, x∈ {r, g, b}, (r r, r g, r b) = (255,0,0), (g r, g g, g b) = (0,255, 0), (b _r , b _g , b _b ) = (0, 0, 255))
The distance between a and b is calculated by Equation 2,

Red is defined as a color chart P _r of the red in the captured image the average value of the red channel in in the area where there is a color chart, color processing method of in green and blue captured image of the color chart of is also the same, P _g and P _b can be obtained,
In step S22, each pixel in the image is adjusted for the red, green, and blue channels, and the red channel value in the region where the red color chart is present approaches 255, thereby processing the green and blue color charts. Is the same,
_{f r = 255 / P r,} f g = 255 / P g, when a f _b = 255 / P _b, the red channel C _r of each pixel in the image by adjusting the _C'r = f _r × C _r the green channel C _g and the blue channel C _b of each pixel by adjusting the _C'g = f _g × C _g and _{C'b = f b × C b} , C'r, C'g, RGB of _C'b Color information is acquired, The jaundice measuring system of Claim 2 characterized by the above-mentioned. The _Nx value is set to correspond to the resolution of the camera,
The jaundice measurement system according to claim 4, wherein a preferable color threshold of the color chart is 10. In step S31, after converting to HSV color information using RGB color information, the facial skin of the image is within the facial contour range E in the captured image based on the hue range of the set face pixel parameter. The area pixel set SK is selected and the total number _NSK thereof is calculated, and the jaundice area pixel set J of the image is further selected from the face range pixels based on the pixel parameter of the jaundice area set in step S32.
In step S33, the ratio JSR = N _J / N _SK of the total number N _J and N _SK of the jaundice region pixel set J is obtained, and when the JSR value falls below the set value th _JSR , the quality of the captured image is set to be poor. Recommend re-importing the image,
In step S34, the color saturation average value J _s of the pixel set J is calculated, and then converted into a jaundice index by the function T (x). The obtained jaundice index is used as the jaundice value J _v obtained by the blood test. The jaundice measuring system according to claim 4. In step S311, the total number N _E of the pixels in the contour line and the saturation average value S _E are calculated and searched, and the ratio SER of N _SK and N _E is calculated,
The step S311 calculates a ratio SER = N _SK / N _E between the total number N _SK of the face skin region pixel set SK and the total number N _E of pixel sets E within the face outline range, and the SER value is th _SER. If the value falls below the value, it is judged that the quality of the captured image is poor, and it is recommended that the image is captured again, and when the _SER value falls below the th _SER value, the quality of the captured image is set to be poor. It is recommended to recapture the image, and the saturation average value S _E of _E is obtained. When the S _E value falls below the th _SE value, it is determined that the ambient light is insufficient when the image is captured. jaundice measuring system according to claim 6 which is recommended to re capture images, SER value exceeds the th _SER values, and S _E value, characterized in that the flow proceeds when step S32 that exceeds the th _SE values. The face pixel parameter is 42.5 ± 18.5,
The th _SER value is 70%;
The th _SE value is 0.5;
The pixel parameter of the jaundice is 40-60,
The th _JSR value is 50%;
The system for measuring jaundice according to claim 7, wherein the function T (x) is a regression statistical function of Js and Jv. The system for measuring jaundice according to claim 1, wherein the postnatal time of the newborn is loaded in the photographing process, thereby responding to the determination of the normal range of jaundice of the newborn.

Images