A method and system for evaluating the degree of teeth mark tongue in traditional Chinese tongue diagnosis, and a storage medium

Abstract

The application relates to a tooth mark tongue degree evaluation method in traditional Chinese medicine tongue diagnosis, which comprises the following steps: acquiring a tongue binary image; extracting edge points of a tongue area; calculating convex vertices on the edge points of the tongue area; fitting the convex vertices on the edge points of the tongue area with a curve based on a moving least square method; filling a region surrounded by the curve; subtracting a region where the tongue image is located to obtain a tooth mark region; calculating a fuzzy coefficient index of each tooth mark; and performing an inverse tangent function operation on a sum of indexes of all tooth marks and normalizing.

Description

Technical Field

[0001] The present invention relates to the field of traditional Chinese medicine treatment, and more specifically, to a method, system and storage medium for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis. Background Art

[0002] Tongue diagnosis falls under the category of visual diagnosis in Traditional Chinese Medicine (TCM). It involves visually examining a patient's tongue texture, coating, and morphology to understand internal changes. Scalloped tongue and its severity are key components of tongue diagnosis. Previously, tongue diagnosis was performed through direct observation of the patient's tongue. However, with the continued modernization and objectification of Traditional Chinese Medicine (TCM), coupled with the development of tongue image acquisition equipment, the use of tongue images to analyze scalloped tongue has become increasingly important for the objectification of TCM diagnosis and treatment.

[0003] The analysis of tongue tooth marks is based on the area formed when the tongue does not have tooth marks, and the existing tongue area is subtracted to obtain the difference area, which is considered to be the tooth mark area. The current method for processing tooth marks in tongue images includes convex hull algorithm. For example, the patent "Device and method for tongue tooth mark recognition based on tongue contour line" (application number: CN201811001462.X) proposes a tongue tooth mark recognition method, which calculates the tongue contour convex hull as an approximate tongue tooth mark contour line based on the tongue contour line, and then determines the tooth mark position based on the maximum distance between the tongue contour line and the approximate tongue tooth mark contour line. Whether it is a tooth mark is determined based on whether the maximum distance is greater than a predetermined threshold. This method uses the convex hull as the tooth mark contour line for a normal tongue, which has a large deviation in contour. The threshold for judging whether it is a tooth mark is difficult to determine, and it is not reasonable to use a unified tooth mark threshold for different people with different tongue sizes.

[0004] Currently, in clinical diagnosis and treatment, the evaluation and analysis of the mild, moderate and severe degree of tooth-marked tongue relies on the doctor's judgment. There is also tooth mark analysis based on tongue images. The patent "Quantitative Analysis Method of Tongue Texture Based on Multi-scale Convolutional Neural Network" (Application Number: CN201910470107.5) imitates the doctor's tooth-marked tongue diagnosis process, and uses the product of tooth mark depth and tooth mark number as an indicator to divide tooth-marked tongue into four categories: no tooth marks, mild tooth marks, moderate tooth marks and severe tooth marks. The degree is based on the value of the tooth mark index, but it cannot fully reflect the characteristics of the tooth marks based solely on tooth mark depth information. Selecting the number of tooth marks as the evaluation criterion in the analysis process is difficult to accurately portray the actual situation of tooth-marked tongue.

[0005] Existing methods use the polygonal area of ​​the convex hull to fit a pre-set normal tongue, and then subtract the tooth mark area from the tongue image area. This method has large errors and is difficult to accurately extract the tooth mark area of ​​the tongue. Due to the differences in the depth of tooth marks, it is difficult to determine the standard for calculating the number of tooth marks. The method of using the number of tooth mark areas to represent the characteristics of the tooth-marked tongue and to judge the degree of tooth marks is unreasonable. The current evaluation standard for the degree of tooth-marked tongue is unclear, and the tooth mark analysis method of tongue images is not entirely reasonable in analyzing the tooth mark area and calculating the number of tooth marks. As a result, the evaluation of the degree of tooth-marked tongue is not objective and accurate, which affects the application of tongue-image-based tooth-marked tongue in traditional Chinese medicine diagnosis and treatment. Summary of the Invention

[0006] The present invention addresses the problem that the existing method of polygonal fitting of the edge of the tongue area to preset a normal tongue does not conform to the edge features of the tongue body, and that some areas of the tongue body are concave due to the influence of the teeth. The moving least squares method is used to fit the edge closed curve of the preset normal tongue body. The area formed by the closed curve minus the existing tongue image area can be used to obtain the tongue's tooth mark area, so as to more realistically reflect the tooth mark condition of the tongue body. At the same time, in response to the defects of the existing technology lacking an evaluation standard for the degree of tooth-marked tongue and not utilizing the objectivity and standardization of traditional Chinese medicine diagnosis and treatment, a tooth-marked tongue evaluation index method based on a fuzzy strategy is proposed. The size and depth of each tooth-marked area are comprehensively considered for each tooth-marked area, and a fuzzy coefficient index for each tooth-marked degree is proposed. The fuzzy coefficient indexes of all tooth-marked degrees of the tongue body are then accumulated, and then transformed and normalized to form an evaluation index for the degree of tooth-marked tongue.

[0007] In order to achieve the above invention objectives, the technical solution adopted is:

[0008] A method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis comprises the following steps:

[0009] Obtain a tongue binary image;

[0010] Extract edge points of the tongue area;

[0011] Calculate the convex vertices of the edge points of the tongue area;

[0012] Based on the moving least squares method, a curve is used to fit the convex vertices on the edge points of the tongue area;

[0013] Fill the area enclosed by the curve;

[0014] Subtract the area where the tongue image is located to obtain the tooth mark area;

[0015] Calculate the fuzzy coefficient index of each tooth mark;

[0016] The sum of all tooth mark indices is subjected to an inverse tangent function operation and normalized.

[0017] Preferably, the specific process of extracting the edge points of the tongue area is as follows: among the points with a value of 1 and the largest column value in the tongue binary image, any one is selected as P1, and the next edge point P2 is searched from the adjacent points of the point, and the search for the next edge point is continued until the starting point P1 is found. The above points are stored in the array Points in sequence, and P1 is stored in the array Points again as the last item.

[0018] Preferably, the specific process of calculating the convex vertex of the edge point of the tongue area is as follows:

[0019] Set the Vertex array to store the convex vertices on the edge points of the tongue area;

[0020] Put the starting point P1 in Points into the Vertex array;

[0021] For any point P in the Points array i , let the two adjacent points be P i-1 and P i+1 , calculate the line segment P i- 1P i and P i P i+1 The cross product of crossValue:

[0022] crossValue=(P i-1 (x)-P i (x))*(P i+1 (y)-P i (y))-(P i-1 (y)-P i (y))*

[0023] (P i+1 (x)-P i (x))

[0024] If crossValue is greater than 0, then point P i It is not a convex vertex on the edge of the tongue area, otherwise it is a convex vertex on the edge of the tongue area and is added to the array Vertex;

[0025] Traverse all points on the edge point array Points in order, and add the convex vertices on the edge points of the tongue area to the Vertex array in order.

[0026] Preferably, the specific process of using a curve to fit the convex vertices of the edge points of the tongue region based on the moving least squares method is as follows:

[0027] The fitting function is expressed as follows:

[0028]

[0029] Where a(x)=(a1(x),a2(x),…,a n (x)) T is the coefficient vector, q(x)=(q1(x),q2(x),…,q n (x)) T is the basis function;

[0030] At each point x, choose an appropriate a(x) so that u h (x) can well approximate u(x); define the discrete weighted approximation function L 2 The norm is used to measure the degree of approximation of a function:

[0031]

[0032] Where N is the number of discrete points in the support domain, r is the radius of the divided support domain, w(|xx i | / r) is a weighting function whose value varies with x and x i The distance between s=|xx i | decreases with an increase;

[0033] Formula (2) is written in matrix form as follows:

[0034] J=(Qa(x)-u) T W(x)(Qa(x)-u) (3)

[0035] In the formula

[0036] is the matrix form of the basis function

[0037] u=[u(x1),u(x2),…,u(x n )] T is the matrix form of the fitting function

[0038] W(x)=diag(w1(s),w2(s),…,w n (s)) is the matrix form of the weighting function

[0039] Based on the least squares principle, we can obtain:

[0040] a(x)=A -1 (x)B(x)u (4)

[0041] Where A(x) = Q T W(x)Q, B(x)=Q T W(x)

[0042] Substituting equation (4) into equation (1) yields the fitting function of function u(x); thereby obtaining the fitting curve of the convex vertices on the edge point, filling the area contained in the curve with 1, and setting the value of the remaining area to 0, thereby obtaining the preset regional binary image of the normal tongue.

[0043] Preferably, the specific process of calculating the fuzzy coefficient index of each tooth mark is as follows:

[0044] Assume that the tooth mark binary image g has N tooth mark regions. For the i-th tooth mark, let the area of ​​the tooth mark region be A i , the side length of the tooth mark fitting curve is S i and height H i , where S i It is obtained by the number of edge points of the tooth mark fitting curve; for each point of the tooth mark curve, the farthest distance of the tooth mark area is calculated by emitting a ray perpendicular to the tangent of the tooth mark edge curve. The farthest distance of all edge points is the height of the tooth mark H. i ;

[0045] For the i-th tooth mark, its size index α i The ratio of the area of ​​the tooth mark to the area of ​​the reference tooth mark is used to obtain the value. The larger the tooth mark area, the more serious the tooth mark. If α i >1, then α i The value of is set to 1;

[0046] For the i-th tooth mark, its tooth mark depth index β i , which is obtained by the ratio of the tooth mark's height to its outer edge length and the ratio of the reference tooth mark's height to its side length. The calculation formula is: If β i >1, then β i The value of is set to 1;

[0047] Since the severity of each tooth mark is determined by the size and depth of the tooth mark, the fuzziness index ω for the i-th tooth mark is i =α i ×β i .

[0048] Preferably, the specific process of performing the arc tangent function operation on the sum of all tooth mark indices and normalizing the result is as follows:

[0049] The fuzziness index ω of all tooth mark areas i Add them together to get the tooth mark tongue degree index Perform inverse tangent operation on ω and normalize it to get It is an indicator of the severity of scalloped tongue.

[0050] At the same time, the present invention also provides a tooth-marked tongue degree evaluation system in traditional Chinese medicine tongue diagnosis, including a tongue binary image acquisition module, an edge point extraction module, a vertex calculation module, a vertex fitting module, a region filling module, a tooth-marked region acquisition module, a fuzzy coefficient index calculation module, and a tooth-marked tongue degree index calculation module;

[0051] Wherein, the tongue binary image acquisition module is used to acquire the tongue binary image;

[0052] The edge point extraction module is used to extract the edge points of the tongue area;

[0053] The vertex calculation module is used to calculate the convex vertices of the edge points of the tongue area;

[0054] The vertex fitting module is used to fit the convex vertices on the edge points of the tongue area with a curve based on the moving least squares method;

[0055] The area filling module is used to fill the area enclosed by the curve;

[0056] The tooth mark area acquisition module is used to obtain the tooth mark area by subtracting the area where the tongue image is located;

[0057] The fuzzy coefficient index calculation module is used to calculate the fuzzy coefficient index of each tooth mark;

[0058] The tooth mark tongue degree index calculation module is used to perform an inverse tangent function operation on the sum of all tooth mark indices and normalize it.

[0059] In addition, the present invention also provides a storage medium, which stores a program therein, and when the program is run, the above-mentioned method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis is executed.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The present invention addresses the problem that the existing method of polygonal fitting of the edge of the tongue area to preset a normal tongue does not conform to the edge features of the tongue body, and that some areas of the tongue body are concave due to the influence of the teeth. The moving least squares method is used to fit the edge closed curve of the preset normal tongue body. The area formed by the closed curve minus the existing tongue image area can be used to obtain the tongue's tooth mark area, so as to more realistically reflect the tooth mark condition of the tongue body. At the same time, in response to the defects of the existing technology lacking an evaluation standard for the degree of tooth-marked tongue and not utilizing the objectivity and standardization of traditional Chinese medicine diagnosis and treatment, a tooth-marked tongue evaluation index method based on a fuzzy strategy is proposed. The size and depth of each tooth-marked area are comprehensively considered for each tooth-marked area, and a fuzzy coefficient index for each tooth-marked degree is proposed. The fuzzy coefficient indexes of all tooth-marked degrees of the tongue body are then accumulated, and then transformed and normalized to form an evaluation index for the degree of tooth-marked tongue. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a flow chart of the method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis. DETAILED DESCRIPTION

[0063] The accompanying drawings are for illustrative purposes only and are not to be construed as limiting this patent;

[0064] The present invention is further described below with reference to the accompanying drawings and embodiments.

[0065] Example 1

[0066] like Figure 1 As shown, the method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis provided by the present invention comprises the following steps:

[0067] (1) Using the segmented tongue binary image f, the value of the tongue part is 1 and the value of the background part is 0;

[0068] (2) Extract the edge points of the tongue area. In the tongue binary image f, find the bottom point, that is, the point with a value of 1 and the largest column value in the two-dimensional image. Select any one of them as P1, search for the next edge point P2 from the adjacent points of this point, and continue searching for the next edge point Pi until the starting point P1 is found. The above points are stored in the array Points in order, and P1 is stored in the array Points again as the last item.

[0069] (3) Search all points in the array Points to obtain the convex vertices on the edge of the tongue area:

[0070] 1) Set the Vertex array to save the convex edge vertices;

[0071] 2) From step (2), we can see that P1 is the starting point P1 in Points placed in the Vertex array;

[0072] 3) For any point P in the Points array i , let the two adjacent points be P i-1 and P i+1 , calculate the line segment P i-1 P i and P i P i+1 The cross product of crossValue:

[0073] crossValue=(P i-1 (x)-P i (x))*(P i+1 (y)-P i (y))-(P i-1 (y)-P i (y))*

[0074] (P i+1 (x)-P i (x))

[0075] 4) If crossValue is greater than 0, then point P i If it is not a point on the convex hull of the tongue edge, otherwise it is a point on the convex hull and is added to the array Vertex;

[0076] 5) Traverse all points on the edge point array Points in order, and add the edge convex hull vertices to the Vertex array in order;

[0077] (4) For the discrete point set formed by all points in the Vertex array, a curve is fitted to the vertices with convex edges based on the moving least squares method:

[0078] 1) The fitting function is expressed as follows:

[0079]

[0080] Where a(x)=(a1(x),a2(x),…,a n (x)) T is the coefficient vector, q(x)=(q1(x),q2(x),…,q n (x)) T As the basis function, preferably a quadratic basis

[0081] q(x)=(1,x,y,x 2 ,xy,y 2 ) T

[0082] 2) At each point x, choose an appropriate a(x) so that u h (x) can well approximate u(x). Define the discrete weighted approximation function L 2 The norm measures the degree of approximation of a function:

[0083]

[0084] Where N is the number of discrete points in the support domain, r is the radius of the divided support domain, w(|xx i | / r) is a weighting function whose value varies with x and x i The distance between s=|xx i |decreases with the increase of |.

[0085] 3) Formula (2) is written in matrix form as follows:

[0086] J=(Qa(x)-u) T W(x)(Qa(x)-u) (3)

[0087] In the formula

[0088] is the matrix form of the basis function

[0089] u=[u(x1),u(x2),…,u(x n )] T is the matrix form of the fitting function

[0090] W(x)=diag(w1(s),w2(s),…,w n (s)) is the matrix form of the weighting function

[0091] 4) Based on the least squares principle, we can obtain:

[0092] 5) a(x) = A -1 (x)B(x)u (4)

[0093] Where A(x) = Q T W(x)Q, B(x)=Q T W(x)

[0094] 6) Substitute equation (4) into equation (1) to obtain the fitting function of function u(x). This can be used to obtain the fitting curve of the vertices of the edge convex hull. The area included in the filled curve is set to 1, and the rest of the area is set to 0, thus obtaining the preset normal tongue area binary image f ‘ .

[0095] (5) Calculate the tooth mark area, using the preset normal tongue area image f ′ Subtract the current tongue area image f to obtain the binary image g = f ‘ -f.

[0096] (6) Calculate the side length of the tongue area in the binary image f, excluding the tongue root part, and use L to represent it. Let a variable C represent the number of tooth marks on the tongue. Considering that the number of normal human teeth is between 28 and 32, the range of C can be selected from 14 to 16. It is preferably set to C = 15. The base length of the reference tooth mark S = L / C can be obtained. The deepest tooth mark is set to an equilateral triangle, and the area of ​​the equilateral triangle with a side length of S is Use this as a reference for the area of ​​the tooth marks.

[0097] (7) Assume that the tooth mark binary image g has N tooth mark areas, analyze the characteristics of each tooth mark area, and calculate the fuzzy coefficient index of each tooth mark degree

[0098] 1) For the i-th tooth mark, let the area of ​​the tooth mark be A i , the side length of the tooth mark fitting curve is S i and height H i , where S iIt can be obtained by the number of edge points of the tooth mark fitting curve; for each point of the tooth mark curve, the farthest distance of the tooth mark area can be calculated by emitting a ray perpendicular to the tangent of the tooth mark edge curve. The farthest distance of all edge points is the height of the tooth mark H. i .

[0099] 2) For the i-th tooth mark, its size index α i The ratio of the area of ​​the tooth mark to the area of ​​the reference tooth mark is used to obtain the value. The larger the tooth mark area, the more serious the tooth mark. If α i >1, then α i is set to 1.

[0100] 3) For the i-th tooth mark, its tooth mark depth index β i , which is obtained by the ratio of the tooth mark's height to its outer edge length and the ratio of the reference tooth mark's height to its side length. The calculation formula is: If β i >1, then β i is set to 1.

[0101] 4) Since the severity of each tooth mark is determined by the size and depth of the tooth mark, the fuzziness index ω for the i-th tooth mark is i =α i ×β i , from the above formula, we can see that for the tooth mark area where the tooth mark is too small and the depth is very shallow, the index ω i The value of will become very small, and the evaluation of the entire tooth-marked tongue will have little effect, which is consistent with the actual situation.

[0102] 5) Set the fuzziness index ω of all tooth mark areas i Add them together to get the tooth mark tongue degree index Since the value of ω may exceed 1, it is preferred to perform an inverse tangent function operation on ω and normalize it to obtain It is an indicator of the severity of scalloped tongue, with a value between 0 and 1, reflecting the condition of scalloped tongue truly and accurately.

[0103] Obviously, the above embodiments of the present invention are merely examples for the purpose of clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the claims of the present invention.

Claims

1. A method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis, characterized by: The following steps are involved: Obtain a tongue binary image; Extract edge points of the tongue area; Calculate the convex vertices of the edge points of the tongue area; Based on the moving least squares method, a curve is used to fit the convex vertices on the edge points of the tongue area; Fill the area enclosed by the curve; Subtract the area where the tongue image is located to obtain the tooth mark area; Calculate the fuzzy coefficient index of each tooth mark; Perform arc tangent function operation on the sum of all tooth mark indices and normalize them; The specific process of extracting the edge points of the tongue area is as follows: among the points with a value of 1 and the largest column value in the tongue binary image, one is randomly selected as P1, and the next edge point P2 is searched from the adjacent points of the point, and the search for the next edge point is continued until the starting point P1 is found. The above points are stored in the array Points in order, and P1 is stored in the array Points again as the last item; The specific process of calculating the convex vertex of the edge point of the tongue area is as follows: Set the Vertex array to store the convex vertices on the edge points of the tongue area; Put the starting point P1 in Points into the Vertex array; For any point P in the Points array i , let the two adjacent points be P i-1 and P i+1 , calculate the line segment P i-1 P i and P i P i+1 The cross product of crossValue: crossValue=(P i-1 (x)-P i (x))*(P i+1 (y)-P i (y))-(P i-1 (y)-P i (y))*(P i+1 (x)-P i (x)) If crossValue is greater than 0, then point P i It is not a convex vertex on the edge of the tongue area, otherwise it is a convex vertex on the edge of the tongue area and is added to the array Vertex; Traverse all points on the edge point array Points in order, and add the convex vertices on the edge points of the tongue area to the Vertex array in order; The specific process of calculating the fuzzy coefficient index of each tooth mark is as follows: Assume that the tooth mark binary image g has N tooth mark regions. For the i-th tooth mark, let the area of ​​the tooth mark region be A i , the side length of the tooth mark fitting curve is S i and height H i , where S i It is obtained by the number of edge points of the tooth mark fitting curve; for each point of the tooth mark curve, the farthest distance of the tooth mark area is calculated by emitting a ray perpendicular to the tangent of the tooth mark edge curve. The farthest distance of all edge points is the height of the tooth mark H. i ; For the i-th tooth mark, its size index α i The ratio of the area of ​​the tooth mark to the area of ​​the reference tooth mark is used to obtain the value. The larger the tooth mark area, the more serious the tooth mark. If α i >1, then α i The value of is set to 1; For the i-th tooth mark, its tooth mark depth index β i , which is obtained by the ratio of the tooth mark's height to its outer edge length and the ratio of the reference tooth mark's height to its side length. The calculation formula is: If β i >1, then β i The value of is set to 1; Since the severity of each tooth mark is determined by the size and depth of the tooth mark, the fuzziness index ω for the i-th tooth mark is i =α i ×β i .

2. The method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis according to claim 1, characterized in that: The specific process of using a curve to fit the convex vertices of the edge points of the tongue area based on the moving least squares method is as follows: The fitting function is expressed as follows: Where a(x)=(a1(x),a2(x),…,a n (x)) T is the coefficient vector, q(x)=(q1(x),q2(x),…,q n (x) T is the basis function; At each point x, choose an appropriate a(x) so that u h (x) can well approximate u(x); define the discrete weighted approximation function L 2 The norm is used to measure the degree of approximation of a function: Where N is the number of discrete points in the support domain, r is the radius of the divided support domain, w(|xx i | / r) is a weighting function whose value varies with x and x i The distance between s=|xx i | decreases with an increase; Formula (2) is written in matrix form as follows: J=(Qa(x)-u) T W(x)(Qa(x)-u) (3) In the formula is the matrix form of the basis function u=[u(x1),u(x2),…,u(x n )] T is the matrix form of the fitting function W(x)=diag(w1(s),w2(s),…,w n (s)) is the matrix form of the weighting function Based on the least squares principle, we can obtain: a(x)=A -1 (x)B(x)u (4) where A(x) = Q T W(x)Q, B(x) = Q T W(x) Substituting equation (4) into equation (1) yields the fitting function of function u(x); thereby obtaining the fitting curve of the convex vertices on the edge point, filling the area contained in the curve with 1, and setting the value of the remaining area to 0, thereby obtaining the preset regional binary image of the normal tongue.

3. A system for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis, characterized by: It includes a tongue binary image acquisition module, an edge point extraction module, a vertex calculation module, a vertex fitting module, a region filling module, a tooth mark region acquisition module, a fuzzy coefficient index calculation module, and a tooth mark tongue degree index calculation module; Wherein, the tongue binary image acquisition module is used to acquire the tongue binary image; The edge point extraction module is used to extract the edge points of the tongue area; The vertex calculation module is used to calculate the convex vertices of the edge points of the tongue area; The vertex fitting module is used to fit the convex vertices on the edge points of the tongue area with a curve based on the moving least squares method; The area filling module is used to fill the area enclosed by the curve; The tooth mark area acquisition module is used to obtain the tooth mark area by subtracting the area where the tongue image is located; The fuzzy coefficient index calculation module is used to calculate the fuzzy coefficient index of each tooth mark; The tooth mark tongue degree index calculation module is used to perform an inverse tangent function operation on the sum of all tooth mark indices and normalize the result; The specific process of extracting the edge points of the tongue area is as follows: among the points with a value of 1 and the largest column value in the tongue binary image, one is randomly selected as P1, and the next edge point P2 is searched from the adjacent points of the point, and the search for the next edge point is continued until the starting point P1 is found. The above points are stored in the array Points in order, and P1 is stored in the array Points again as the last item; The specific process of calculating the convex vertex of the edge point of the tongue area is as follows: Set the Vertex array to store the convex vertices on the edge points of the tongue area; Put the starting point P1 in Points into the Vertex array; For any point P in the Points array i , let the two adjacent points be P i-1 and P i+1 , calculate the line segment P i-1 P i and P i P i+1 The cross product of crossValue: crossValue=(P i-1 (x)-P i (x))*(P i+1 (y)-P i (y))-(P i-1 (y)-P i (y))*(P i+1 (x)-P i (x)) If crossValue is greater than 0, then point P i It is not a convex vertex on the edge of the tongue area, otherwise it is a convex vertex on the edge of the tongue area and is added to the array Vertex; Traverse all points on the edge point array Points in order, and add the convex vertices on the edge points of the tongue area to the Vertex array in order; The specific process of calculating the fuzzy coefficient index of each tooth mark is as follows: Assume that the tooth mark binary image g has N tooth mark regions. For the i-th tooth mark, let the area of ​​the tooth mark region be A i , the side length of the tooth mark fitting curve is S i and height H i , where S i It is obtained by the number of edge points of the tooth mark fitting curve; for each point of the tooth mark curve, the farthest distance of the tooth mark area is calculated by emitting a ray perpendicular to the tangent of the tooth mark edge curve. The farthest distance of all edge points is the height of the tooth mark H. i ; For the i-th tooth mark, its size index α i The ratio of the area of ​​the tooth mark to the area of ​​the reference tooth mark is used to obtain the value. The larger the tooth mark area, the more serious the tooth mark. If α i >1, then α i The value of is set to 1; For the i-th tooth mark, its tooth mark depth index β i , which is obtained by the ratio of the tooth mark's height to its outer edge length and the ratio of the reference tooth mark's height to its side length. The calculation formula is: If β i >1, then β i The value of is set to 1; Since the severity of each tooth mark is determined by the size and depth of the tooth mark, the fuzziness index ω for the i-th tooth mark is i =α i ×β i .

4. A storage medium having a program stored therein, characterized in that: When the program is running, the method for evaluating the degree of tooth-marked tongue in traditional Chinese medicine tongue diagnosis according to any one of claims 1 to 2 is executed.

Images