Preparation method of color block, colorimetric card for evaluating skin diseases and application

Abstract

The invention provides a preparation method of color blocks in a colorimetric card for evaluating skin diseases, the colorimetric card for evaluating skin diseases and application, and the method comprises the steps: taking the skin color of a real skin disease patient as a reference, and obtaining different types of color blocks through a computer algorithm of a deep convolutional neural network; the similarity of the colorimetric card is closer to that of an actual skin damage color, and the distance between two continuous chromaticity is closer to the color difference of the actual skin damage color. Furthermore, the colorimetric card prepared by the method can be used for evaluating skin diseases, especially psoriasis.

Description

Technical Field

[0001] The invention belongs to the technical field of medical appliances, and in particular relates to a method for preparing a color block, a colorimetric card for evaluating skin diseases using the color block, and a use thereof. Background Art

[0002] The skin is the largest organ in the human body, and skin diseases are a general term for diseases that occur in the skin and skin appendages. In my country, there are a large number of patients in dermatology clinics. There are many types of skin diseases. Clinically, many skin diseases have similar symptoms, and accurate diagnosis of the type of skin disease is crucial to treatment. Accurate assessment of the color of skin lesions is very important for the diagnosis and treatment of most skin diseases. For example, red macules are more common in wind-heat and blood-heat, white macules are more common in blood deficiency and wind noise, and cyanotic macules are more common in blood stasis. For another example, the ABCDE rule is often used in the diagnosis of malignant melanoma, where C represents color judgment. When the skin lesions show brown, black, red, white or blue pigmentation diffusion, or the color is significantly different from the color of other moles of the patient, the probability of disease increases. For example, psoriasis is an immune-mediated chronic, recurrent, inflammatory, systemic disease induced by the combined effects of genetics and the environment. The typical clinical manifestations are scaly erythema or plaques. The Psoriasis Area and Severity Index (PASI) is currently commonly used to assess the severity of psoriasis. An important factor in this assessment system is the degree of erythema of the diseased skin lesions, which is generally divided into "light red", "red", "dark red" and "extremely dark red", and are assigned 1 to 4 points respectively.

[0003] At present, the clinical evaluation of the color of skin lesions is mostly based on the naked eye, lacking quantitative standards. Doctors, nurses or patients themselves mostly rely on experience when evaluating the skin lesions. If the evaluator is experienced, the evaluation result will be more objective. If the experience is insufficient, there will be a large error in the evaluation result, which will affect the diagnosis and treatment of the patient's condition.

[0004] In the prior art, one way to improve the reliability of clinical evaluation data of skin diseases is to design a standard colorimetric card for comparison and reference of skin lesion colors. The oldest colorimetric card is the Munsel colorimetric card, which uses two indicators, hue and saturation, to define each color. Although these colorimetric cards provide many useful data for evaluators, they lack similarity with the color of normal skin lesions, and the interval between two consecutive chromaticities is too large, resulting in unsatisfactory evaluation results of the colorimetric cards.

[0005] The utility model patent with application number 202320229242.2 discloses a medical colorimetric card, which is specially used to distinguish the bile inside the drainage bag used after the treatment of hepatobiliary patients. The structure of this colorimetric card is relatively complex and is not suitable for the color assessment of skin diseases. The invention patent application with application number 202211473007.6 discloses a colorimetric card and its remote comparison method. The colorimetric card is provided with a color comparison area for color comparison with the real object on both side arms of the U-shaped bottom plate, and a number of color blocks are evenly arranged in the color comparison area. The real object comparison area for revealing the real object to be compared is between the two side arms of the U-shaped bottom plate. This colorimetric card is particularly suitable for comparing the tongue. When used for the skin, due to the different sizes of the upper and lower openings of the U-shaped bottom plate, it will interfere with the skin color, and the color blocks on the two side arms are asymmetrical, so there will be a large error when comparing the skin colors on both sides. The invention patent application with application number 201410260673.0 discloses a color card and a method and system for testing skin, wherein the color card can be a circular or square color ring, and the color blocks are symmetrically arranged. The invention patent application with application number 202310671545.4 discloses a quantification diagnosis system for pigmented skin disease lesion area based on a neural network algorithm, which introduces the neural network algorithm into the skin disease diagnosis system.

[0006] The invention patent applications with application numbers 201410260673.0 and 202310671545.4 both focus on post-processing of images. However, in order to conveniently, quickly and accurately evaluate the color of skin lesions of skin diseases, a colorimetric card with a closer similarity to the actual color of skin lesions and a spacing between two consecutive chromaticities closer to the color difference of the actual color of skin lesions is the key. If the previous skin lesion color evaluation is inaccurate, then no matter how advanced the post-processing technology is, it will be useless. The key to the colorimetric card is the proximity of the color block to the actual color of the skin lesions. The closer the color block is to the actual color of the skin lesions, the more accurate the evaluation of the skin disease will be. Summary of the invention

[0007] In the current field of skin disease lesion color assessment, there is an urgent need for a set of color blocks that are closer to the actual skin lesion color, so as to prepare a more convenient and quick colorimetric card. It can be used for unified, objective and accurate assessment of the color of skin disease lesions.

[0008] In view of the above problems existing in the prior art, the present invention aims to provide a method for preparing color blocks in a colorimetric card for evaluating skin diseases and a corresponding colorimetric card. The present invention takes the skin color of real skin disease patients as a benchmark and utilizes a computer algorithm of a deep convolutional neural network to obtain color blocks of different categories, thereby producing a colorimetric card that is closer to the actual skin lesion color and the distance between two continuous chromaticities is closer to the color difference of the actual skin lesion color.

[0009] Specifically, the present invention provides the following technical solutions.

[0010] In one aspect, the present invention provides a method for preparing a color block in a colorimetric card for evaluating skin diseases, and the preparation steps are as follows:

[0011] Step 01: Take photos of the skin lesions of the skin disease patients according to a unified shooting standard to obtain an image set of the affected skin parts;

[0012] Step 02: Segment the skin lesion area one by one for the above-mentioned image sets to obtain the precise target skin lesion area;

[0013] Step 03: Perform color sampling on the segmented target skin lesion area image;

[0014] Step 04: Cluster the sampling results to obtain the centroid color value of each cluster;

[0015] Step 05: Create color blocks based on the obtained centroid color values.

[0016] According to the above preparation method, the unified shooting standard described in step 01 includes that the light source is set to D65 light source, the shooting device is a SLR camera or a smart phone, the sensitivity (ISO) is set to 100 to 200, the number of pixels is not less than 2 million, and the picture format is JPEG, PNG or BMP.

[0017] Preferably, according to the above preparation method, step 01 comprises the following steps:

[0018] Step 011: Clean the patient's skin at the affected area and remove any surface substances that may affect the shooting effect (such as ointments or cosmetics).

[0019] Step 012: Use a fixing device (such as a tripod) to fix the shooting equipment to avoid shooting quality problems (such as unstable focus or blurred images) caused by shaking of the equipment when taking pictures.

[0020] Step 013: Take photos at an angle perpendicular to the surface of the lesion;

[0021] More preferably, according to the above preparation method, the skin lesion area segmentation in step 02 is automated by computer algorithm, with I∈N w×h×3 represents a RGB three-channel skin image with a pixel resolution of w×h, I ij =(r,g,b) represents the pixel value of the i-th row and j-th column; M∈{0,1} w×h It represents the mask obtained by segmenting image I, and its scale is the same as that of image I. ij∈{0,1} represents the value of the i-th row and j-th column of the mask, M ij =1 means that the pixel at this position in image I belongs to the target skin lesion area, M ij =0 means that the pixel at this position in image I does not belong to the target skin lesion area.

[0022] More preferably, according to the above preparation method, step 03 comprises the following steps:

[0023] Step 031: Convert the r, g, b values ​​of each pixel in image I into L*a*b* color values. The converted image is represented by T∈N w×h×3 , T ij =(L, a, b) represents the L*a*b* color value of the pixel in the i-th row and j-th column;

[0024] Step 032: Use sliding window W∈{1} m×n The segmented target lesion area is sampled, where m represents the width of the window and n represents the height of the window.

[0025] More preferably, according to the above preparation method, wherein in step 032, T ij When sampling the pixel position, the sliding window covers the area from the i-th to (i+m-1)-th rows to the j-th to (j+n-1)-th columns. If all pixels in the window area belong to the target lesion area, the color values ​​of all pixels in the window are calculated for sampling, otherwise the position is not operated. The window sliding direction is first from left to right along the image width direction, and then from top to bottom along the image height direction. The sliding step length is L pixels.

[0026] More preferably, according to the above preparation method, the sampling algorithm adopts the arithmetic mean, that is, U ij =∑ i≤u<i+m-1,j≤v<j+n-1 T u,v / (m×n).

[0027] More preferably, according to the above preparation method, the size of the window W may be 3×3 pixels, 5×5 pixels or 7×7 pixels.

[0028] More preferably, according to the above preparation method, L is set to 3 pixels.

[0029] More preferably, according to the above preparation method, step 04 is performed as follows: all the values ​​U sampled in step 032 are ij Sequence U = [u1,u2,u3…u n ], the sequence length is n,u i=(L, a, b), the sequence U can be regarded as a series of points in the Lab three-dimensional color space; cluster the data sequence U to obtain K clusters, and the centroid C of each cluster i =(L,a,b),i∈[1,K],C i represents a color, K represents the number of target colors obtained after clustering, and the color set C = {C i} represents the typical color distribution of skin lesions in a real-world population of skin disease patients.

[0030] More preferably, according to the above preparation method, wherein K is set to a multiple of 4.

[0031] More preferably, according to the above preparation method, wherein K is set to 8.

[0032] More preferably, according to the above preparation method, step 05 comprises the following steps:

[0033] Step 051: Color sequence C i Sort by C i The L values ​​are sorted from large to small and marked with serial numbers. The color sequence after sorting is S i , i is the color identification number, i∈[1,n];

[0034] Step 052: Make color blocks: Set the color sequence S i Print one by one to form color blocks.

[0035] On the other hand, the present invention also provides a colorimetric card for evaluating skin diseases, which includes a colorimetric card plate, a color block prepared according to the above preparation method, a first positioning color block, a second positioning color block and a hollow area.

[0036] Preferably, according to the above colorimetric card, the hollow area is a hollowed-out area in the center of the colorimetric card plate.

[0037] More preferably, according to the above colorimetric card, the colorimetric card plate is a rectangular frame or a square frame.

[0038] More preferably, according to the above colorimetric card, the color block is a rectangular frame or a square block.

[0039] More preferably, according to the above-mentioned colorimetric card, a first positioning color block of the same color is respectively arranged at the upper left corner and the lower left corner of the colorimetric card board, a second positioning color block of the same color is respectively arranged at the upper right corner and the lower right corner of the colorimetric card board, the colors of the first positioning color block and the second positioning color block are different, and the colors of the first positioning color block and the second positioning color block are both different from the colors of the color blocks.

[0040] More preferably, according to the above-mentioned color chart, color blocks are arranged between the first positioning color block and the second positioning color block on the upper and lower borders of the color chart board, each color block has the same size, and the arrangement order of the color blocks on the lower border is the same as the arrangement order of the color blocks on the upper border and is symmetrically arranged.

[0041] More preferably, according to the above colorimetric card, it also includes rectangular handles on the left and right sides of the colorimetric card plate, which have the same size and are symmetrically distributed on the left and right sides.

[0042] More preferably, according to the above colorimetric card, a circular suction cup is attached to the lower surface of the two handles, with the opening of the cup facing downward, for adsorbing the colorimetric card plate to the patient's skin surface.

[0043] On the other hand, the present invention also provides a use of the above-mentioned colorimetric card in evaluating skin diseases, characterized in that a colorimetric card of appropriate size is selected according to the size of the patient's skin lesions, and the colorimetric card is attached to the patient's skin lesion site to ensure that the entire skin lesion is located in the hollow area of ​​the colorimetric plate; the evaluator compares the color of the patient's skin lesions with the colorimetric card to obtain the most appropriate color match.

[0044] Preferably, according to the above use, the skin disease is psoriasis.

[0045] The method of the present invention uses the skin color of real skin disease patients as a reference, and uses a computer algorithm of a deep convolutional neural network to obtain color blocks of different categories, thereby producing a colorimetric card that is closer to the actual skin lesion color and the distance between two consecutive chromaticities is closer to the color difference of the actual skin lesion color. Compared with the prior art, the following improved technical effects are achieved:

[0046] 1. In the present invention, real-world data is obtained by sampling and clustering the colors of skin lesions of a large number of real patients. The extracted colors are used as the colors of color blocks, which are more in line with clinical practice needs and can significantly improve the accuracy of skin disease assessment;

[0047] 2. In the present invention, the same color block is deployed in the upper and lower frames of the color card and is symmetrically distributed. By selecting colorimetric plates of different sizes, the hollow area of ​​the colorimetric plate can be tightly surrounded by the entire skin lesion, shortening the distance between the skin lesion and the color block, which helps to improve the accuracy of colorimetry;

[0048] 3. A circular suction cup is attached to the bottom of the handle of the colorimetric card of the present invention, so that the card can be attached to the skin surface very conveniently during colorimetric comparison, thereby improving the convenience of operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the colorimetric card structure of the present invention;

[0050] Figure 2is a schematic diagram of color sampling using a sliding window in the present invention;

[0051] Figure 3 It is a schematic flow chart of the method for making a colorimetric card in the present invention;

[0052] Among them, 1-colorimetric card board; 2-color block; 3-first positioning color block; 4-second positioning color block; 5-hollow area; 6-handle; 7-suction cup. DETAILED DESCRIPTION

[0053] In order to explain the present invention more clearly and make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention will be described in detail with reference to the accompanying drawings and detailed implementation methods. Obviously, the drawings described below are only some implementation methods of the present invention. For ordinary technicians in this field, other implementation methods can be obtained based on these drawings without paying creative labor.

[0054] Specifically, the specific implementation modes of the present invention are described in detail below with reference to the accompanying drawings.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the principles of the present invention. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0056] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention in detail, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0057] First, the colorimetric card for evaluating skin diseases of the present invention will be described.

[0058] like Figure 1As shown, the colorimetric card plate 1 is a rectangular frame, and the hollowed-out part in the middle is a hollow area 5; there is a square first positioning color block 3 at the upper left corner and the lower left corner of the rectangular frame, the color is the same and different from the color of the color block 2; there is a square second positioning color block 4 at the upper right corner and the lower right corner of the rectangular frame, the color is the same and different from the color of the color block 2 and the color of the first positioning color block 3 at the upper left corner and the lower left corner; the color block 2 is arranged between the two first positioning color blocks 3 and the second positioning color blocks 4 on the left and right sides of the upper frame of the rectangular frame, and the color block 2 is rectangular Or square, each color block 2 has the same size; the color block 2 is arranged between the two first positioning color blocks 3 and the second positioning color blocks 4 on the left and right sides of the lower frame of the rectangular frame, and the arrangement order of the color blocks 2 is the same as the arrangement order of the color blocks 2 in the upper frame; that is, the color blocks 2 at the same positions of the upper and lower frames of the colorimetric card board 1 have the same color, and the left and right sides of the colorimetric card board 1 are respectively rectangular handles 6 with the same size and symmetrical distribution on the left and right. A circular suction cup 7 is pasted on the lower surface of the two handles 6, with the disc opening facing downward, for adsorbing the colorimetric card board 1 on the patient's skin surface.

[0059] In this preferred embodiment, the size of the colorimetric card plate 1 is 6 cm*5 cm, the size of the hollow area 5 is 4 cm*3 cm, the size of the handle 6 is 2 cm*1.5 cm, and the positioning color blocks 3, 4 and the color block 2 are all squares with a side length of 0.5 cm.

[0060] Preferably, when evaluating the color of psoriasis lesions, there are four common descriptions of the color of psoriasis lesions in clinical practice, namely "light red", "red", "dark red" and "extremely dark red". Therefore, in this embodiment, the upper frame of the rectangular frame is provided with 8 color blocks 2, which are arranged from left to right in the order of light to dark red tones; the lower frame of the rectangular frame is also provided with 8 color blocks 2, and the colors and arrangement order are the same as those of the upper frame.

[0061] Reference below Figure 2 and 3 The preparation method of the color block 2 used in the colorimetric card of the present invention and the preparation method of the colorimetric card are described in detail. For the convenience of description, psoriasis, which is a common skin disease, is used as an example for explanation. In clinical practice, there are four common descriptions of the color of psoriasis lesions, namely "light red", "red", "dark red" and "very dark red". In the Psoriasis Area and Severity Index (PASI) score, 1-4 points are assigned respectively, and the higher the score, the lower the score represents the severity of the disease.

[0062] The specific steps for preparing the color blocks in the color chart used to evaluate skin diseases are as follows:

[0063] Step 01: Take photos of the skin lesions of the skin disease patients according to a unified shooting standard to obtain an image set of the affected skin parts;

[0064] In the specific implementation, the patient group consisted of 50 psoriasis patients who visited our hospital, aged from 20 to 70 years old, with psoriasis severity ranging from mild, moderate to severe. A Canon D800 SLR camera was used to take pictures. When taking pictures, the shooting equipment was fixed on a tripod, the flash was forced to be turned on, and the shooting direction was perpendicular to the surface of the lesion. The picture size was 1500*2000, and the picture format was JPEG.

[0065] Step 02: Segment the skin lesion area one by one for the above-mentioned image sets to obtain an accurate target skin lesion area ROI (Region of Interest);

[0066] I∈N w×h×3 represents an RGB three-channel skin image with a pixel resolution of w×h, I ij =(r,g,b) represents the pixel value of the i-th row and j-th column. The skin lesion area segmentation work described in step 02 can be done manually or automatically using a computer algorithm to obtain the accurate pixel area of ​​the diseased skin lesion.

[0067] In this embodiment, a computer algorithm based on a deep convolutional neural network is used to segment the color image to obtain a mask M∈{0,1} having the same width and height as the image I. w×h , whose scale is the same as that of image I, M ij ∈{0,1} represents the value of the i-th row and j-th column of the mask, M ij =1 means that the pixel at this position in image I belongs to the target skin lesion area, M ij =0 means that the pixel at this position in image I does not belong to the target skin lesion area.

[0068] Step 03: Perform color sampling on the segmented target skin lesion area image;

[0069] First, the r, g, b color values ​​of each pixel in image I are converted into CIELab color values. The converted image is represented by T∈N w×h×3 , T ij =(L, a, b) represents the pixel value of the i-th row and j-th column;

[0070] Use sliding window W∈{1} m×n The target lesion area obtained by segmentation is sampled, where m represents the width of the window and n represents the height of the window; ijTaking the pixel as an example, when sampling the position, the sliding window covers the area of ​​the i-th to (i+m-1)-th rows and the j-th to (j+n-1)-th columns. If all pixels in the window area belong to the target lesion area, the sampling is valid. The arithmetic mean of the color values ​​of all pixels in the window is calculated, which is expressed as U ij =∑ i≤u<i+m-1,j≤v<j+n-1 T u,v / (m×n), as the sampling result; otherwise, the position is not sampled;

[0071] Preferably, the scale of the window W can be 3×3 pixels, 5×5 pixels, or 7×7 pixels. More preferably, the sliding window W is set to 5×5 pixels. The larger the size of the original image, the more conducive it is to reduce the influence of noise points by using a larger window. The window sliding direction is from left to right and from top to bottom.

[0072] Preferably, the sliding step size L is set to 3 pixels.

[0073] Step 04: Cluster the sampling results to obtain the centroid color value of each cluster;

[0074] All the color values ​​U sampled in step 03 ij Composition sequence U = [u1,u2,u3…u n ], the sequence length is n,u i =(L, a, b), the sequence U can be regarded as a series of points in the three-dimensional color space of L, a, b, representing the color space distribution of skin lesions in the psoriasis patient group in the real world;

[0075] Cluster the data sequence U;

[0076] Preferably, K-means clustering is used in this embodiment; since there are generally four descriptions of the color of psoriasis lesions, namely "light red", "red", "dark red" and "very dark red", the number of clusters is generally a multiple of 4, such as 8 or 12. In this embodiment, the number of categories is set to 8, resulting in 8 clusters, and the centroid of each cluster is C i =(L,a,b),i∈[1,8],C i represents a color; color set C = {C i} represents the typical color categories of skin lesions in the psoriasis patient population.

[0077] Step 05: Based on the obtained centroid color value, a colorimetric card device is made, which can be widely used in the color assessment of skin lesions of skin diseases.

[0078] Step 051: Color sequence C i Sort by C iThe L values ​​are sorted from large to small and marked with serial numbers. The color sequence after sorting is S i , i is the color identification number, i∈[1,8]; preferably, referring to the commonly used descriptions of psoriasis lesion colors including "light red", "red", "dark red", and "very dark red", the color sequence S i , divided into 4 groups according to the order from light to dark red, with 2 colors in each group, S1 and S2 are classified as light red, S3 and S4 are classified as red, S5 and S6 are classified as dark red, and S7 and S8 are classified as very dark red; each color is printed into a color block 2 with a size of 0.5 cm × 0.5 cm;

[0079] Step 052: Make a colorimetric card; in this embodiment, the first positioning color block 3, the second positioning color block 4 and the color block 2 are all squares with a side length of 0.5 cm, the size of the colorimetric card plate 1 is 6 cm*5 cm, the size of the hollow area 5 is 4 cm*3 cm, and the size of the handle 6 is 2 cm*1.5 cm;

[0080] The color sequence S i Print one by one, each color forms a square color block 2, which is pasted between the first positioning color block 3 and the second positioning color block 4 on the left and right sides of the upper frame of the rectangular frame in order from left to right, and each color block 2 has the same size; between the first positioning color block 3 and the second positioning color block 4 on the left and right sides of the lower frame of the rectangular frame, the arrangement order of the color blocks 2 is the same as that of the upper frame; the size of the hollow area 5 can be freely set according to the common skin lesion size of different skin diseases;

[0081] Preferably, the colors of the two square-shaped first positioning color blocks 3 at the upper left corner and the lower left corner of the rectangular frame are set to white, and the colors of the two square-shaped second positioning color blocks 4 at the upper right corner and the lower right corner of the rectangular frame are set to black;

[0082] When using the colorimetric card of the present invention to evaluate skin lesions, the colorimetric card is attached to the patient's skin lesion site to ensure that the entire skin lesion is located in the hollow area 5 of the colorimetric plate; the evaluator compares the color of the patient's skin lesion with the colorimetric card to obtain the most suitable color match; or the same shooting standard as step 01 is used to take pictures of the above skin lesions and the colorimetric card and archive them.

[0083] The above is a specific implementation scheme of the present invention. Without violating the spirit and principle of the present invention, some structures and details may be changed or modified, and these changes or modifications are within the protection scope of the present invention.

Claims

1. A method for preparing color blocks in a colorimetric card for evaluating skin diseases, the preparation steps are as follows: Step 01: Take photos of the skin lesions of the skin disease patients according to a unified shooting standard to obtain an image set of the affected skin parts; Step 02: Segment the skin lesion area one by one for the above-mentioned image sets to obtain the precise target skin lesion area; Step 03: Perform color sampling on the segmented target skin lesion area image; Step 04: Cluster the sampling results to obtain the centroid color value of each cluster; Step 05: Create color blocks based on the obtained centroid color values.

2. The preparation method according to claim 1, wherein The unified shooting standards described in step 01 include setting the light source to D65 light source, using a SLR camera or a smart phone as the shooting device, setting the sensitivity (ISO) to 100 to 200, the number of pixels to be no less than 2 million, and the image format to JPEG, PNG or BMP.

3. The preparation method according to claim 1 or 2, wherein Step 01 includes the following steps: Step 011: Clean the skin of the patient's skin lesions to remove the surface substances that may affect the shooting effect; Step 012: Use a fixing device to fix the shooting device to avoid shooting quality problems caused by shaking of the device; Step 013: Take pictures at an angle perpendicular to the surface of the lesion.

4. The preparation method according to any one of claims 1 to 3, wherein: The lesion area segmentation described in step 02 is automated using a computer algorithm, with I∈N w×h×3 represents a RGB three-channel skin image with a pixel resolution of w×h, I ij =(r,g,b) represents the pixel value of the i-th row and j-th column; M∈{0,1} w×h It represents the mask obtained by segmenting image I, and its scale is the same as that of image I. ij ∈{0,1} represents the value of the i-th row and j-th column of the mask, M ij =1 means that the pixel at this position in image I belongs to the target skin lesion area, M ij =0 means that the pixel at this position in image I does not belong to the target skin lesion area.

5. The preparation method according to any one of claims 1 to 4, wherein: Step 03 includes the following steps: Step 031: Convert the r, g, b values ​​of each pixel in image I into L*a*b* color values. The converted image is represented by T∈N w×h×3 , T ij =(L, a, b) represents the L*a*b* color value of the pixel in the i-th row and j-th column; Step 032: Use sliding window W∈{1} m×n The segmented target lesion area is sampled, where m represents the width of the window and n represents the height of the window; preferably, the scale of the window W is 3×3 pixels, 5×5 pixels or 7×7 pixels.

6. The preparation method according to claim 5, wherein: In step 032, T ij When sampling the position of a pixel, the sliding window covers the area of ​​the i-th to (i+m-1)-th rows and the j-th to (j+n-1)-th columns. If all pixels in the window area belong to the target lesion area, the color values ​​of all pixels in the window are calculated for sampling, otherwise the position is not operated. The sliding direction of the window is first from left to right along the width of the image, and then from top to bottom along the height of the image. The sliding step is L pixels, preferably L is set to 3 pixels. The preferred sampling algorithm adopts the arithmetic mean, that is, U ij =∑ i≤u<i+m-1,j≤v<j+n-1 T u,v / (m×n)。 7. The preparation method according to claim 5 or 6, wherein: Step 04 is as follows: all the values ​​U sampled in step 032 are ij Sequence U = [u1,u2,u3…u n ], the sequence length is n,u i =(L, a, b), the sequence U can be regarded as a series of points in the Lab three-dimensional color space; cluster the data sequence U to obtain K clusters, and the centroid C of each cluster i =(L,a,b),i∈[1,K],C i represents a color, K represents the number of target colors obtained after clustering, and the color set C = {C i } represents the typical color distribution of skin lesions in the real world among patients with skin diseases; Preferably, K is set to a multiple of 4, preferably 8.

8. The preparation method according to claim 7, wherein: Step 05 includes the following steps: Step 051: Color sequence C i Sort by C i The L values ​​are sorted from large to small and marked with serial numbers. The color sequence after sorting is S i , i is the color identification number, i∈[1,n]; Step 052: Make color blocks: Set the color sequence S i Print one by one to form color blocks.

9. A colorimetric card for evaluating skin diseases, comprising a colorimetric card plate, a color block prepared according to the preparation method according to any one of claims 1 to 8, a first positioning color block, a second positioning color block and a hollow area.

10. The colorimetric card according to claim 9, wherein: The hollow area is the area hollowed out in the center of the colorimetric card plate; Preferably, the colorimetric card plate is a rectangular frame or a square frame, and / or the color blocks are rectangular blocks or square blocks; More preferably, a first positioning color block of the same color is respectively arranged at the upper left corner and the lower left corner of the color comparison card board, and a second positioning color block of the same color is respectively arranged at the upper right corner and the lower right corner of the color comparison card board, the colors of the first positioning color block and the second positioning color block are different, and the colors of the first positioning color block and the second positioning color block are both different from the colors of the color blocks; Further preferably, color blocks are arranged between the first positioning color block and the second positioning color block on the upper and lower frames of the color comparison card board, each color block has the same size, and the arrangement order of the color blocks on the lower frame is the same as the arrangement order of the color blocks on the upper frame and is symmetrically arranged.

11. The colorimetric card according to claim 9 or 10, further comprising rectangular handles on the left and right sides of the colorimetric card plate, which are of the same size and symmetrically distributed on the left and right sides; preferably, a circular suction cup is attached to the lower surface of each handle, with the opening facing downward, for adsorbing the colorimetric card plate to the patient's skin surface.

12. Use of the colorimetric card according to any one of claims 9 to 11 in evaluating skin diseases, characterized in that: According to the size of the patient's skin lesions, choose a colorimetric card of appropriate size and stick it on the patient's skin lesions to ensure that the entire lesion is located in the hollow area of ​​the colorimetric card; the evaluator compares the color of the patient's skin lesions with the colorimetric card to obtain the most appropriate color match.

13. The use according to claim 12, wherein The skin disease is psoriasis.

Images