A method and system for generating a medical report of the severity of a skin burn

Abstract

The application discloses a kind of medical report generation methods and systems of skin burn severity, comprising: the obvious edge texture of burn area is accurately obtained by edge detection technology, which helps to accurately define the scope of burn.By judging the pixel point in edge texture, and carrying out contour fitting, the overall edge of burn area is obtained, so that the doctor can more intuitively understand the burn situation.The use of downsampling and region segmentation technology not only simplifies the data processing process, but also realizes the clear distinction of different burn degree areas.By calculating the proportion of each burn degree area, the doctor is provided with objective and accurate burn severity evaluation basis.In addition, data statistics are carried out on the local area of different burn degree, and detailed medical report is generated, which provides strong support for the diagnosis and treatment of doctors, and provides scientific basis for the rehabilitation and treatment plan of patients.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image processing, and relates to a medical report generation method and system for skin burn severity. BACKGROUND

[0002] As a global public health problem, the assessment and treatment of skin burns face many challenges. Burn not only has high mortality and disability rate, but also its prognosis and treatment plan are highly dependent on the accurate assessment of burn area and depth. However, the traditional burn degree assessment method mainly depends on the observation of the appearance and area size of the patient's burn wound by the clinician, which has strong subjectivity and low accuracy, especially in the area estimation of large area burn patients.

[0003] In order to more accurately assess the burn area and depth, computer-aided methods are introduced into the precise measurement of the burn field. These technologies can provide more objective and standardized judgment methods through image processing, which can help generate more scientific burn severity medical reports. In image processing technology, mean shift algorithm is an effective tool that can effectively capture burn areas with complex shape and uneven gray scale distribution by iteratively updating the cluster center. However, the mean shift algorithm iterates multiple times for each pixel point, which may cause obvious time complexity problem in the clinical scene that needs to generate medical reports quickly.

[0004] Therefore, it is an urgent problem to be solved in the current burn treatment field to innovatively improve the medical report generation process and use image processing technology to more accurately and objectively judge the severity of burn wounds. SUMMARY

[0005] The purpose of the present application is to solve the problem of time complexity caused by the mean shift algorithm which iterates multiple times for each pixel point in the prior art, and to provide a medical report generation method and system for skin burn severity.

[0006] To achieve the above purpose, the technical scheme is adopted as follows:

[0007] A medical report generation method for skin burn severity, comprising:

[0008] Collecting the image of the burn site of the patient and preprocessing the image of the burn site of the patient;

[0009] Performing edge detection on the preprocessed burn site image to obtain the obvious edge texture of the burn area;

[0010] determine whether the pixel points contained in the obvious edge texture of the burn area are the pixel points of the overall edge of the burn area, if yes, then perform contour fitting on the obtained overall edge pixel points to obtain the overall edge of the burn area;

[0011] down-sample the burn area, and segment the down-sampled burn area to obtain the areas where different burn degrees are located;

[0012] statistically burn data based on the local areas of different burn degrees to obtain a medical report of the skin burn severity.

[0013] Further improvements of the present application are as follows:

[0014] Further, edge detection is performed on the pre-processed burn site image to obtain the obvious edge texture of the burn area, specifically: based on the canny operator, edge detection is performed on the pre-processed burn site image to obtain the obvious edge texture in the image, the edge texture includes the limb edge and the edge texture of the burn area, and the edge texture of the burn area includes the internal edge and the peripheral edge, the appearance of the internal edge indicates that there are burn areas with different severity degrees in the whole burn area; the limb edge is directly detected by edge detection.

[0015] Further, before determining whether the pixel points contained in the obvious edge texture of the burn area are the pixel points of the overall edge of the burn area, it further includes: performing maximum circumscribed rectangle on the edge texture of the burn area, based on the positional relationship between the coordinate position of the edge texture pixel point and the coordinate position of the center point of the rectangle, and the gradient value of the edge texture pixel point, obtaining the overall edge coefficient of the pixel points in the edge texture in the circumscribed rectangle, specifically:

[0016]

[0017] wherein, A i represents the overall edge coefficient of the pixel point i in the edge texture in the rectangle; (x, y) represents the coordinate of the center point of the rectangle; (x i ,y i ) represents the coordinate of the i-th edge texture pixel point; L i represents the shortest distance between the i-th edge texture pixel point and the circumscribed rectangle; T i represents the gradient value of the i-th edge texture pixel point;

[0018] The judgment of whether the pixel points contained in the obvious edge texture of the burn area are the pixel points of the overall edge of the burn area is specifically: based on the relationship between the overall edge coefficient and the set threshold value, it is judged whether the pixel points in the edge texture in the circumscribed rectangle belong to the overall edge pixel points, if yes, all the obtained overall edge pixel points are profile fitted to obtain the overall edge of the burn area.

[0019] Further, the down-sampled burn area is segmented to obtain the area where different burn degrees are located, specifically:

[0020] Based on the mean shift algorithm, the pixel points in the down-sampled image are iterated to construct a density coefficient.

[0021] Based on the otsu algorithm idea, the density coefficient histogram is analyzed to obtain a local area where a density center is suspected to exist;

[0022] The obtained local area is regarded as a positioning area, and the total number of pixel points in different positioning areas and the density coefficient in the positioning area are obtained to obtain the bandwidth variation coefficient of the positioning area;

[0023] Based on the bandwidth variation coefficient of the positioning area, the bandwidth size of the positioning area is obtained; the initial bandwidth parameter size of the non-positioning area is set, the mean shift algorithm is performed on the non-positioning area, and the bandwidth size of the non-positioning area is obtained;

[0024] Based on the bandwidth parameters of the mean shift algorithm in different areas, the positioning of the local area with different burn degrees is completed.

[0025] Further, based on the mean shift algorithm, the pixel points in the down-sampled image are iterated to construct a density coefficient, specifically:

[0026] Based on the mean shift algorithm, all the pixel points in the down-sampled image are iterated for the first time, and after iteration, the pixel points will move from the original position to the to-be-determined density center after the first iteration, and the distance moved in the middle is the first iteration distance, G j G represents the iteration distance of the jth movement, J represents the maximum iteration number, and the area where the density center is located is positioned by the change of the pixel point iteration distance;

[0027]

[0028] Wherein, B represents the density coefficient of the pixel point in the down-sampled image, and each pixel point will correspond to a coefficient; G init G represents the distance moved from the initial position to the stop position when the pixel point itself is iterated for the first time, that is, the first iteration distance; J represents the maximum iteration number; G j G represents the iteration distance of the jth iteration of the pixel point; G init,kThe first iteration distance of the kth pixel point in the 8-neighborhood around the pixel point;

[0029] The density coefficient histogram is analyzed based on the otsu algorithm to obtain a local region in which a density center is suspected to exist, specifically: the gray histogram of the preprocessed gray image is converted into a density coefficient histogram, the density coefficient histogram is analyzed based on otsu threshold segmentation, all possible threshold values are traversed, and the inter-class variance under each threshold value is calculated, finally the threshold value that maximizes the inter-class variance is selected as the segmentation threshold, and the image is segmented into two parts of foreground and background by the segmentation threshold, wherein the foreground part is the region in which a density center is suspected to exist.

[0030] Further, the bandwidth variation coefficient of the positioning region is obtained based on the total number of pixel points in different positioning regions and the density coefficient in the positioning region, specifically:

[0031]

[0032] Wherein, C represents the bandwidth variation coefficient of the positioning region, each bandwidth region corresponds to a coefficient; P represents the number of positioning regions; Q p represents the total number of pixel points in the pth positioning region; B p,q represents the density coefficient of the qth pixel point in the pth positioning region; Max{} represents taking the maximum value in the list.

[0033] Further, the bandwidth size of the positioning region is obtained based on the bandwidth variation coefficient of the positioning region, specifically:

[0034] Set R init as the initial bandwidth parameter size, the initial bandwidth parameter R init is used in the non-positioning region to perform the mean shift algorithm to obtain the bandwidth size of the non-positioning region; while in the positioning region, the bandwidth size of the pth positioning region is R p , and Wherein represents the upward rounding symbol.

[0035] Furthermore, based on the bandwidth parameters of the mean-shift algorithm applied to different regions, the localization of areas with different degrees of burns is completed. Specifically, the corresponding positions of the localized areas in the downsampled image are restored in the original image to obtain the original localized areas. For each pixel in the original localized area, the weighted average of all points in its neighborhood is calculated based on the neighborhood range defined by the bandwidth parameters, with the weights determined by the Gaussian kernel. The pixel is then moved to a new position, and the above drifting process is repeated until the point converges to a density center. This density center is a clustering center, and all pixels that converge to the same density center are divided into the same segmentation region. Finally, all pixels in the image are divided into different segmentation regions, thus completing the image segmentation task.

[0036] Furthermore, based on the statistical analysis of localized burn data at different burn degrees, a medical report on the severity of skin burns is obtained, specifically:

[0037] Based on the proportion of the total burn area in the burn site and the proportion of local areas with different burn degrees in the total burn area, data on the proportion of the overall and local burn area are obtained.

[0038] The average gray value of normal area pixels in the burned area is used as the skin color of the normal area. The greater the difference between the gray value of the skin color in different local burned areas and the gray value of the skin color in the normal area, the more severe the burn. The severity of the burn in the local area constitutes the severity of the burn in the overall burned area, thus obtaining the data on the severity of the burn in the local burned areas of the overall burn.

[0039] Determine whether the burned area crosses a joint. If not, no additional annotation is needed in the image; if so, additional note is needed in the medical report.

[0040] By combining the patient's personal information, burn site image data, locally segmented burn areas marked in the burn site image, and additional information such as severity data and whether it crosses multiple joints, a medical report on the severity of the patient's skin burns can be obtained, and a corresponding report can be generated.

[0041] A medical report generation system for the severity of skin burns, comprising:

[0042] The preprocessing module acquires images of the patient's burn sites and performs preprocessing on these images.

[0043] An edge detection module performs edge detection on the preprocessed image of the burn area to obtain the obvious edge texture of the burn area.

[0044] A judgment module judges whether the pixel points contained in the obvious edge texture of the burn area are the pixel points of the overall edge of the burn area, and if so, performs contour fitting on the obtained overall edge pixel points to obtain the overall edge of the burn area.

[0045] A segmentation module performs down-sampling on the burn area and segments the down-sampled burn area to obtain the areas where different burn degrees are located.

[0046] An acquisition module acquires the medical report of the skin burn severity based on the burn data of the local areas of different burn degrees.

[0047] Compared with the prior art, the present application has the following beneficial effects:

[0048] The present application accurately obtains the obvious edge texture of the burn area through edge detection technology, which helps to accurately define the burn range. By judging the pixel points in the edge texture and performing contour fitting, the overall edge of the burn area is obtained, so that the doctor can more intuitively understand the burn situation. The use of down-sampling and area segmentation technology not only simplifies the data processing process, but also realizes the clear distinction of different burn degree areas. By calculating the proportion of each burn degree area, the doctor is provided with objective and accurate burn severity evaluation basis. In addition, data statistics are performed on the local areas of different burn degrees to generate detailed medical reports, which provide strong support for the diagnosis and treatment of doctors and provide scientific basis for the rehabilitation and treatment plan of patients. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0050] Figure 1 The flowchart of the method for generating the medical report of the skin burn severity of the present application;

[0051] Figure 2 The structure diagram of the medical report generation system of the skin burn severity of the present application;

[0052] Figure 3 The schematic diagram of iterative movement. DETAILED DESCRIPTION

[0053] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some but not all of the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations.

[0054] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0055] It should be noted that: similar reference numbers and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0056] In the description of the embodiments of the present application, it should be noted that, if the orientation or position relationship indicated by the terms "upper", "lower", "horizontal", "inner" and the like is based on the orientation or position relationship shown in the drawings, or is the orientation or position relationship when the product of the present application is usually placed, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, therefore, it cannot be understood as a limitation on the present application. In addition, the terms "first", "second" and the like are only used for differentiation in description, and cannot be understood as indicating or implying relative importance.

[0057] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly inclined. For example, "horizontal" only means that its direction is relatively more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly inclined.

[0058] In the description of the embodiments of the present application, it should also be noted that, unless otherwise explicitly specified and limited, if the terms "arrangement", "installation", "connection", "connection" appear, they should be understood in a broad sense, for example, can be fixedly connected, can be detachably connected, or integrally connected; can be mechanically connected, can be electrically connected; can be directly connected, can be indirectly connected through an intermediate medium, or can be the communication inside two elements. For those of ordinary skill in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0059] The present application will be described in further detail below with reference to the drawings:

[0060] Referring toFigure 1 The application discloses a medical report generation method for skin burn severity, comprising:

[0061] S101, collecting a burn site image of a patient and preprocessing the burn site image of the patient;

[0062] The preprocessing of the burn site image of the patient is specifically: converting the collected burn site image into a gray-scale image, and obtaining a corresponding gray-scale value of each pixel point of the color image through a weighted average method, so as to generate the gray-scale image.

[0063] Suppose that the collected original color image is I(x, y, c), wherein x and y respectively represent pixel coordinates of the image, and c represents a color channel, and the gray-scale image G(x, y) is:

[0064] G(x, y) = 0.299 * I(x, y, 1) + 0.587 * I(x, y, 2) + 0.114 * I(x, y, 3)

[0065] Wherein, c = 1, 2, 3 respectively correspond to red, green and blue channels.

[0066] S102, performing edge detection on the preprocessed burn site image to obtain obvious edge texture of the burn area;

[0067] The edge detection is performed on the preprocessed burn site image based on a canny operator to obtain obvious edge texture in the image, and the edge texture includes limb edges and edge texture of the burn area, and the edge texture of the burn area includes internal edges and peripheral edges, the appearance of the internal edges indicates that there are burn areas with different severities in the whole burn area; the limb edges are directly detected through edge detection.

[0068] S103, judging whether the pixel points contained in the obvious edge texture of the burn area are pixel points of the whole edge of the burn area, if yes, performing contour fitting on the obtained whole edge pixel points to obtain the whole edge of the burn area;

[0069] Before judging whether the pixel points contained in the obvious edge texture of the burn area are pixel points of the whole edge of the burn area, it further includes: performing a maximum circumscribed rectangle on the edge texture of the burn area, obtaining a whole edge coefficient of the pixel points in the edge texture in the circumscribed rectangle based on a positional relationship between a coordinate position of the edge texture pixel points and a coordinate position of a center point of the rectangle and a gradient value of the edge texture pixel points, and the specific process is as follows:

[0070]

[0071] Wherein, A iEi represents the overall edge coefficient of the pixel point i in the edge texture in the rectangle; (x, y) represents the center point coordinates of the rectangle; (x i ,y i ) represents the coordinates of the i-th edge texture pixel point; L i represents the shortest distance between the i-th edge texture pixel point and the circumscribed rectangle; T i represents the gradient value of the i-th edge texture pixel point;

[0072] The judgment of whether the pixel point contained in the obvious edge texture of the burn area is the pixel point of the overall edge of the burn area is specifically: based on the relationship between the overall edge coefficient and the set threshold value, it is judged whether the pixel point in the edge texture in the circumscribed rectangle belongs to the overall edge pixel point, if yes, all the obtained overall edge pixel points are fitted to obtain the overall edge of the burn area.

[0073] S104, down-sampling the burn area, and segmenting the down-sampled burn area to obtain the area where different burn degrees are located;

[0074] S104.1, based on the mean shift algorithm, iteratively processing the pixel points in the down-sampled image to construct a density coefficient;

[0075] Based on the mean shift algorithm, all the pixel points in the down-sampled image are iterated for the first time, and after iteration, the pixel points will move from the original position to the to-be-determined density center after the first iteration, and the distance moved in the middle is the first iteration distance, G j represents the iteration distance of the j-th movement, J represents the maximum iteration number, and the region where the density center is located is positioned by the change of the iteration distance of the pixel point;

[0076]

[0077] Wherein, B represents the density coefficient of the pixel point in the down-sampled image, and each pixel point will correspond to a coefficient; G init represents the distance moved by the pixel point itself from the initial position to the stop position when the first iteration is performed, that is, the first iteration distance; J represents the maximum iteration number; G j represents the iteration distance obtained by the j-th iteration of the pixel point; G init,k represents the first iteration distance of the k-th pixel point in the 8-neighborhood around the pixel point;

[0078] S104.2, based on the otsu algorithm idea, analyzing the density coefficient histogram to obtain a local area where the density center is suspected to exist;

[0079] The gray scale histogram of the preprocessed gray scale image is converted into a density coefficient histogram, the density coefficient histogram is analyzed based on otsu threshold segmentation, all possible threshold values are traversed, the inter-class variance under each threshold value is calculated, and finally the threshold value that makes the inter-class variance maximum is selected as the segmentation threshold value, the image is segmented into foreground and background two parts by the segmentation threshold value, and the foreground part is a region suspected to have a density center.

[0080] S104.3, regarding the acquired local region as a positioning region, the total number of pixel points in different positioning regions and the density coefficient in the positioning region are acquired to obtain the bandwidth variation coefficient of the positioning region;

[0081]

[0082] Wherein, C represents the bandwidth variation coefficient of the positioning region, each bandwidth region corresponds to a coefficient; P represents the number of positioning regions; Q p represents the total number of pixel points in the pth positioning region; B p,q represents the density coefficient of the qth pixel point in the pth positioning region; Max{} represents taking the maximum value in the list.

[0083] S104.4, based on the bandwidth variation coefficient of the positioning region, the bandwidth size of the positioning region is acquired; the initial bandwidth parameter size of the non-positioning region is set, the mean shift algorithm is performed on the non-positioning region to acquire the bandwidth size of the non-positioning region;

[0084] Set R init as the initial bandwidth parameter size, the initial bandwidth parameter R init is used in the non-positioning region to perform the mean shift algorithm and acquire the bandwidth size of the non-positioning region; in the positioning region, the bandwidth size of the pth positioning region is R p , and Wherein represents the upward rounding symbol.

[0085] S104.5, based on the bandwidth parameter when the mean shift algorithm is performed on different regions, the positioning of the local region with different burn degrees is completed.

[0086] The positioning area in the down-sampling image is restored in the corresponding position in the original image to obtain an original positioning area in the original image, and for a pixel point in the original positioning area, a weighted mean value of all points in the neighborhood of the pixel point is calculated based on the neighborhood range defined by the bandwidth parameter, and the weight is determined by a Gaussian kernel; the pixel point is moved to a new position, and the above drift process is repeated until the point converges to a density center; the density center is a cluster center, and all pixel points converging to the same density center are divided into the same segmentation area; finally, all pixels in the image are divided into different segmentation areas, thereby completing the image segmentation task.

[0087] S105, based on the local area statistics of different burn degrees, the medical report of the skin burn severity is obtained.

[0088] Based on the proportion of the total burn area in the burn site and the proportion of the local area of different burn degrees in the total burn area, the overall and local burn area proportion data is obtained;

[0089] The mean value of the gray value of the normal area pixel point in the burn area is taken as the normal area skin color, and the difference between the gray value of the regional skin color of different local burn areas and the gray value of the normal area skin color is obtained, the greater the difference, the more serious the burn degree, the burn severity of the local area constitutes the burn severity of the overall burn area, and the overall burn local burn area burn severity data is obtained;

[0090] It is judged whether the part with the burn area spans the joint, if not, no additional labeling is needed in the picture, if yes, additional marking is needed in the medical report;

[0091] The patient's personal information, burn site image data, local segmented burn area in the burn site image, and the additional marking information of the severity data and whether it spans multiple joints are unified to obtain the medical report content of the skin burn severity of the corresponding patient, and the corresponding report is generated.

[0092] Referring to Figure 2 The application discloses a kind of medical report generation systems of skin burn severity, comprising:

[0093] The pre-processing module collects the burn site image of the patient, and pre-processes the burn site image of the patient;

[0094] The edge detection module performs edge detection on the pre-processed burn site image to obtain the obvious edge texture of the burn area.

[0095] A judgment module judges whether the pixel points contained in the obvious edge texture of the burn area are pixel points of the overall edge of the burn area, and if so, the obtained overall edge pixel points are subjected to contour fitting to obtain the overall edge of the burn area.

[0096] A segmentation module performs down-sampling on the burn area and segments the down-sampled burn area to obtain the area where different burn degrees are located.

[0097] An acquisition module acquires the medical report of the skin burn severity based on the statistics of the burn data of the local area of different burn degrees.

[0098] Embodiment:

[0099] The application discloses a medical report generation method for skin burn severity, which comprises the following steps:

[0100] Step 1: Obtain the image of the burn site of the patient.

[0101] During the process of obtaining the image of the burn site of the patient, it is necessary to first ensure that the light of the shooting environment is sufficient and uniform, so as to avoid the influence of shadows and overexposure on the image quality. A high-resolution medical imaging device or a multispectral camera is used to shoot the burn area, so as to ensure that the image is clear and the color is true. When shooting, the vertical angle between the camera and the burn site is kept, so as to obtain an accurate two-dimensional image. At the same time, the date, time and basic information of the patient of the shooting are recorded, so as to facilitate the subsequent image preprocessing and analysis and record them as part of the medical report. The obtained image should cover the overall situation of the burn area, including the center and the edge part, so as to ensure the comprehensiveness and accuracy of the subsequent analysis.

[0102] Step 2: Preprocess the obtained burn site image.

[0103] The preprocessing process is to convert the burn site image into a gray-scale image. The RGB value of each pixel point of the color image is calculated by the weighted average method, and the corresponding gray value is replaced by the gray value, so as to generate a gray-scale image.

[0104] Step 3, edge detection is performed on the image after pre-processing to detect obvious edge texture, the overall edge coefficient of the pixel points in the edge texture is defined, the overall edge of the burn area is located by the overall edge coefficient, then the image is standardized, the image resolution is changed and unified, the image is down-sampled to obtain a down-sampled image, the density coefficient of the pixel points in the down-sampled image is defined, and the local area suspected to have a density center is located by the otsu algorithm, the local area located is called a located area, the bandwidth variation coefficient of the located area is defined, the definition of the bandwidth parameter of the mean shift algorithm for different areas is completed, and the local area of different burn degrees is located.

[0105] The canny operator is selected to complete edge detection on the burn site image after pre-processing, the edge detection will detect the obvious edge texture in the image, the detected edge texture includes the edge of the limb and the edge of the burn area, and the edge of the burn area has internal edge and peripheral edge, the appearance of the internal edge indicates that there is a burn area with different severity in the whole burn area.

[0106] Since the pre-processed burn site image has been separated from the limb part, the limb edge can be directly detected by edge detection. The remaining edge texture belongs to the burn area, and the maximum circumscribed rectangle can be made for the remaining edge texture. The pixel points belonging to the overall edge of the burn area are located by the positional relationship of the edge texture in the maximum circumscribed rectangle.

[0107] The overall edge coefficient of the pixel points in the edge texture in the rectangle is defined, and the specific formula is:

[0108]

[0109] Wherein, A i represents the overall edge coefficient of the pixel points in the edge texture in the rectangle; (x,y) represents the center point coordinates of the rectangle; (x i ,y i ) represents the coordinates of the i-th edge texture pixel point; L i represents the shortest distance between the i-th edge texture pixel point and the circumscribed rectangle; T i represents the gradient value of the i-th edge texture pixel point.

[0110] The farther the coordinate position of the edge texture pixel point is from the coordinate position of the rectangular center point, the shorter the shortest distance to the circumscribed rectangle is, and thus the pixel point is more inclined to belong to the overall edge pixel point of the burn, and thus the corresponding overall edge coefficient is larger. In addition, since the overall edge is a separation line between the burn area and the non-burn area, and the non-burn area is normal skin, there is a significant gap in the gray value distribution between the burn area and the non-burn area, and thus the overall edge pixel point can obtain a larger gradient value, and thus the larger the gradient value of the pixel point, the larger the corresponding overall edge coefficient.

[0111] After obtaining the overall edge coefficient A of the pixel point, the overall edge of the burn area is located. The threshold T of the overall edge coefficient A is set A When the A value is greater than the threshold value, it is considered that the pixel point corresponding to the A value belongs to the overall edge pixel point, and here the threshold T A = 0.8, that is, when A > 0.8, it is considered that the pixel point belongs to the overall edge pixel point. The obtained overall edge pixel point is fitted to obtain the overall edge of the burn area.

[0112] After obtaining the overall edge, the range delimited by the overall edge is the burn area on the surface of the current limb of the patient, and thus when the different burn severity areas are divided by the mean shift, the beneficial effects obtained are:

[0113] Only the pixel points in the burn area are subjected to the mean shift algorithm, and the number of pixel points participating in iteration is reduced. When the pixel points in the burn area are iterated to the overall edge, iteration can be stopped, because the range of the burn area is known, and thus the iteration process does not need to be performed in the non-burn area, and the iteration speed of the pixel points in the algorithm process is accelerated.

[0114] After locating the overall edge of the burn area, the internal area surrounded by the edge is the burn area, and because the severity of the burn is different in different local areas, different colors are presented in the burn area, and corresponding to the gray scale image, the gray scale values of the areas with different burn severities are different, and different areas can be segmented by the mean shift algorithm.

[0115] The specific process is as follows:

[0116] The image after the foregoing operation is subjected to a standardization operation, that is, the image is standardized to a 1080P image, and the image resolution is changed to 1920x1080 by using the cv2.resize function to complete the standardization operation.

[0117] The standardized image is subjected to down-sampling, and the specific process is as follows:

[0118] The pixel block with a size of 2*2 is established at the position of pixel coordinate (1, 1), and the non-overlapping pixel blocks with a size of 2*2 are established in the S type to the global image, and finally the number of pixel blocks is 960*540, and the gray value of the pixel block is obtained by using average pooling, that is, the average value of the gray values of the corresponding 2*2 pixel points is used as the gray value of the corresponding pixel block.

[0119] The image obtained after the downsampling is completed is collectively referred to as a downsampling image in the subsequent.

[0120] The beneficial effect of using the downsampling image is that the number of pixel points in the downsampling image is reduced compared with the original image, so that the pixel points can find the density center faster during the iterative operation in the algorithm process.

[0121] The operation object of the mean shift algorithm is a pixel point, and when the mean shift algorithm process is performed on the pixel point in the downsampling image, the result of one iteration is that the pixel point is moved once, as shown in formula (1). Figure 3 The distance of the movement is related to the distance of the pixel point from the density center, and the farther the distance from the density center, the more the number of iterations, and the distance of the movement may be farther, so that the distance of the movement by iteration can reflect the distance relationship between the pixel point and the density center, and in all iterations, the data point is usually located far away from the density center in the first iteration, because when the pixel point is in the initial position, the points in the neighborhood are relatively dispersed, and the weighted average value calculated in the algorithm is different from the position of the current point, so the distance of the movement is larger, but relatively, if the pixel point is close to the density center, the distance of the movement in the first iteration is relatively small.

[0122] The first iteration is performed on all pixel points, and after the iteration, the pixel point is moved from the original position to the density center after the first iteration, and the distance of the movement therebetween is referred to as the first iteration distance in the present application, and the present application uses G j to represent the iteration distance of the jth movement, and J represents the maximum number of iterations, and by the change of the iteration distance of the pixel point, the region where the density center is located can be positioned.

[0123] The density coefficient of the pixel point is:

[0124]

[0125] wherein B represents the density coefficient of the pixel point in the downsampling image, and each pixel point has a corresponding coefficient; G initrepresents the distance of the pixel point itself from the initial position to the stop position in the first iteration, that is, the first iteration distance described in the foregoing; J represents the maximum number of iterations, which is a variable user-defined parameter, and here J = 5; G j represents the iteration distance of the pixel point in the jth iteration; G init,k represents the first iteration distance of the kth pixel point in the 8-neighborhood of the pixel point.

[0126] The density coefficient of the pixel point is embodied by the iteration movement of the pixel point itself and the iteration movement of the 8-neighborhood pixel points around the pixel point.

[0127] Formula The part represents the sum of the first iteration distance of the pixel point itself and the average distance of several subsequent iterations. The greater the result value of this part, the more obvious the iteration movement of the pixel point itself is required to reach the density center, especially the influence of the first iteration distance is more obvious, so the tendency of the pixel point itself as the density center is smaller, and the corresponding B value should be smaller.

[0128] Formula The result of the part represents the sum of the first iteration distances of the pixel points in the 8-neighborhood of the pixel point. The smaller the value, the greater the possibility of the existence of the density center near the neighborhood points. In combination with the previous part of the formula, if the pixel point itself needs to move a large distance to reach the density center, but the distance of the neighborhood pixel points from the density center tends to be shorter, it means that the current pixel point cannot exist as a density center, and the distance relationship between the neighborhood pixel points and the density center is not related to the current pixel point, because the density center to which the current pixel point belongs and the density center to which the neighborhood pixel points belong are not the same density center, so the density coefficient of the current pixel point should be smaller, that is, the formula The smaller the result value of the part, the smaller the B value.

[0129] In actual practice, the pixel point is likely to exist near the edge of a local area with different degrees of burn severity, so the trend of long iteration distance of the pixel point itself and short iteration distance of the neighborhood is likely to occur. The farther away from this trend, the greater the B value is likely to be obtained, and the density center is usually located in the direction away from the edge of the area, which conforms to the logical relationship described above, so it is reasonable that the density coefficient of the pixel point is smaller because of the foregoing logic. Thus, the density coefficient of the pixel point is obtained.

[0130] The greater the density coefficient of the pixel point, the greater the possibility of the pixel point as a density center, so when the density coefficients of the pixel points in a local region are generally large, the local region is likely to have a density center, and therefore the application uses the density coefficients of the pixel points obtained in the foregoing to replace the gray values of the pixel points by using the idea of Otsu threshold segmentation to segment the region suspected of having a density center, and the process is described as follows:

[0131] The gray histogram of the preprocessed gray image is converted into a density coefficient histogram, the density coefficient histogram is analyzed, all possible threshold values are traversed, the inter-class variance under each threshold value is calculated, and finally the threshold value that maximizes the inter-class variance is selected as the segmentation threshold value, and the image is segmented into foreground and background two parts by the segmentation threshold value, wherein the foreground part is the region suspected of having a density center.

[0132] The region to be positioned is referred to as a positioning region, and the number of pixel points and the distribution of the density coefficients of the pixel points of the positioning region are different, and the bandwidth variation coefficient of the positioning region is obtained according to the number of pixel points and the density coefficients of the pixel points of the positioning region.

[0133]

[0134] Wherein, C represents the bandwidth variation coefficient of the positioning region, each bandwidth region corresponds to a coefficient; P represents the number of positioning regions; Q p represents the total number of pixel points of the pth positioning region; B p,q represents the density coefficient of the qth pixel point in the pth positioning region; Max{} represents taking the maximum value in the list.

[0135] In the mean shift algorithm, the bandwidth is used as a key parameter to control the search range of the kernel function. It determines the neighborhood size considered in calculating the kernel density estimate of each data point. If the bandwidth is 5x5, it means that the search and calculation of the kernel function are performed in the 5x5 region around each pixel point. The positioning region and the density coefficients of the pixel points in the positioning region are obtained in the foregoing, the smaller the coverage range of the positioning region, the more accurate the positioning of the density center, and at this time, if the density coefficients of the pixel points are also relatively large, a smaller bandwidth parameter should be used when determining the density center in the region, otherwise, the larger the area of the positioning region, the smaller the distribution of the density coefficients of the pixel points therein, and a larger bandwidth is used. The greater the Q p , the smaller the Q p , and the greater the value of the bandwidth variation coefficient C of the corresponding positioning region.

[0136] After obtaining the bandwidth variation coefficient C value, the bandwidth size of different positioning areas is determined according to the bandwidth variation coefficient of the positioning area. The greater the bandwidth variation coefficient C value, the smaller the obtained bandwidth parameter size, and vice versa. The specific process is defined as follows:

[0137] R init represents the initial bandwidth parameter size. The initial bandwidth parameter R is used in the non-positioning area init The mean shift algorithm is performed, and in the positioning area, the bandwidth size of the pth positioning area is R p , and wherein represents the upward rounding symbol, aiming to keep the bandwidth size as an integer. At this point, the size of the bandwidth parameter when the mean shift algorithm is performed in different areas is obtained.

[0138] The corresponding position of the positioning area in the original image is restored in the down-sampled image, and the original positioning area is obtained in the original image. For the pixel points in the original positioning area, the weighted mean of all points in the neighborhood of each pixel point is calculated based on the neighborhood range defined by the bandwidth parameter, and the weight is determined by the Gaussian kernel. Move the pixel point to a new position and repeat the above drift process until the point converges to a density center. The density center is a cluster center, and all pixel points converging to the same density center are divided into the same segmentation area. Finally, all pixels in the image are divided into different segmentation areas, thereby completing the image segmentation task.

[0139] Step 4, after positioning the local area of different burn degrees, burn-related data is counted, and the generation of the medical report of the skin burn severity is completed.

[0140] After completing the positioning of the local area of different burn degrees, the overall and local burn area ratio data is obtained according to the proportion of the burn total area in the burn site and the proportion of the local area of different burn degrees in the burn total area.

[0141] The skin color of the patient himself will affect the judgment of the burn severity, and the skin color of different areas of the human body is also different. The more severe the burn, the greater the difference between the color of the burn area and the original skin color. Therefore, the average gray value of the normal area pixel points in the burn area is taken as the normal area skin color, and the difference between the gray value of the normal area skin color and the gray value of the different local burn area skin color is calculated. The greater the difference, the more severe the burn. The burn severity of the local area constitutes the burn severity of the overall burn area, and the overall burn local burn area burn severity data is obtained.

[0142] The part where the burn area exists, if not across multiple joints, does not need to be additionally marked, but if the burn area exists across at least two joints and the burn area exists in a connected manner, it needs to be additionally marked in the medical report, because when the burn area crosses multiple joints, it may involve dysfunction of multiple joints, and after the burn, the scar hyperplasia, contracture and skin and tissue damage around the joint will cause limited joint movement, even joint deformity, and the burn across multiple joints needs more complex treatment strategies.

[0143] The patient information, the burn site image data, the local segmentation of the burn area in the burn site image, and the severity data and the additional marking information of whether it crosses multiple joints are unified to obtain the medical report content of the skin burn severity of the corresponding patient, and the corresponding report is generated.

[0144] The above only is the preferred embodiment of the present application, and is not used to limit the present application, for the person skilled in the art, the present application can have various changes and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for generating a medical report on the severity of skin burns, characterized in that, include: Images of the patient's burn sites were acquired and preprocessed. Edge detection is performed on the preprocessed burn area image to obtain the obvious edge texture of the burn area; based on The operator performs edge detection on the preprocessed burn area image to obtain obvious edge textures in the image. The edge textures include the edge textures of the limb edges and the edge textures of the burn area. The edge textures of the burn area include the inner edges and the outer edges. The appearance of the inner edges indicates that there are local burn areas of different severity in the whole burn area. The limb edges are directly detected by edge detection. Determine whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area. If so, perform contour fitting on the obtained overall edge pixels to obtain the overall edge of the burn area. Before determining whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area, the process further includes: constructing a maximum bounding rectangle for the edge texture of the burn area; based on the positional relationship between the coordinates of the edge texture pixels and the coordinates of the center point of the rectangle, and the gradient values ​​of the edge texture pixels, obtaining the overall edge coefficients of the pixels in the edge texture within the bounding rectangle, specifically: ； in, Represents the pixels in the texture of the inner edge of the rectangle. The overall edge coefficient; Indicates the coordinates of the center point of the rectangle; Indicates the first The coordinates of each edge texture pixel; Indicates the first The shortest distance between each edge texture pixel and its bounding rectangle; Indicates the first Gradient values ​​of each edge texture pixel; The determination of whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area is specifically as follows: based on the relationship between the overall edge coefficient and the set threshold, it is determined whether the pixels in the edge texture within the outer rectangle belong to the pixels of the overall edge. If so, the contour of all the obtained overall edge pixels is fitted to obtain the overall edge of the burn area. The burn area was downsampled, and the downsampled burn area was segmented to obtain the areas where different degrees of burns were located. The density coefficients are constructed by iterating through the pixels in the downsampled image using the mean-shift algorithm. based on The algorithm analyzes the density coefficient histogram to identify local regions where density centers may exist. The acquired local area is regarded as the positioning area. The bandwidth variation coefficient of the positioning area is obtained by considering the total number of pixels in different positioning areas and the density coefficient in the positioning area. Based on the bandwidth variation coefficient of the positioning area, the bandwidth size of the positioning area is obtained; the initial bandwidth parameter size of the non-positioning area is set, and the mean drift algorithm is applied to the non-positioning area to obtain the bandwidth size of the non-positioning area. Based on the bandwidth parameters of the mean drift algorithm in different regions, the local area with different burn degrees is located. Based on the statistical analysis of localized burn data of different burn degrees, a medical report on the severity of skin burns is obtained.

2. The method for generating a medical report on the severity of skin burns according to claim 1, characterized in that, The method based on mean shift iterates through the pixels in the downsampled image to construct density coefficients, specifically as follows: The mean-shift algorithm is used to perform the first iteration on all pixels in the downsampled image. After the iteration, the pixels will move from their original positions to the center of the undetermined density after the first iteration, with the distance moved being the k-th ... distance of the next iteration Indicates the first The iteration distance of the next move, using This indicates the maximum number of iterations, and the region where the density center is located is determined by the change in the iteration distance between pixels. ； in, This represents the density coefficient of pixels in the downsampled image; each pixel will have a corresponding coefficient. Indicates the pixel itself. The distance traveled from the initial position to the stopping position during the 1st iteration, i.e., the 1st... Distance at the next iteration; Indicates the maximum number of iterations; Represents the first pixel The iteration distance obtained in the next iteration; Represents the area around a pixel. Within the neighborhood The first pixel Distance at the next iteration; The basis The algorithm analyzes the density coefficient histogram to identify local regions where density centers may exist. Specifically, it converts the grayscale histogram of the preprocessed grayscale image into a density coefficient histogram, based on... Threshold segmentation analyzes the density coefficient histogram, traverses all possible thresholds, and calculates the inter-class variance at each threshold. Finally, the threshold that maximizes the inter-class variance is selected as the segmentation threshold. The image is then segmented into foreground and background parts by the segmentation threshold, with the foreground part being the region where density centers are suspected to exist.

3. The method for generating a medical report on the severity of skin burns according to claim 2, characterized in that, The method of obtaining the bandwidth variation coefficient of the positioning region by considering the total number of pixels in different positioning regions and the density coefficient in the positioning region is as follows: ； in, This represents the bandwidth variation coefficient of the positioning area; each bandwidth area corresponds to one such coefficient. Indicates the number of positioning areas; Indicates the first The total number of pixels in each positioning region; Indicates the first The first location area Density coefficient of each pixel; This indicates that the maximum value is taken from the contents of the list.

4. The method for generating a medical report on the severity of skin burns according to claim 3, characterized in that, The method of obtaining the bandwidth size of the positioning area based on the bandwidth change coefficient of the positioning area is as follows: set up The initial bandwidth parameter size is used in non-location areas. A mean-shift algorithm is used to obtain the bandwidth in the non-localization region; while in the localization region, the first... The bandwidth of each positioning area is ,and ,in The symbol indicates rounding up.

5. The method for generating a medical report on the severity of skin burns according to claim 4, characterized in that, The bandwidth parameter used in the mean-shift algorithm based on different regions is used to locate local areas with different degrees of burns. Specifically, the location area in the downsampled image is restored to its corresponding position in the original image to obtain the original location area. For each pixel in the original location area, a weighted average of all points in its neighborhood is calculated based on the neighborhood range defined by the bandwidth parameter, with the weights determined by a Gaussian kernel. The pixel is moved to the new position, and the mean-shift algorithm is used to iterate over all pixels in the downsampled image. After iteration, the pixel moves from its original position to the undetermined density center after iteration until it converges to a density center. This density center is a clustering center, and all pixels that converge to the same density center are divided into the same segmentation region. Finally, all pixels in the image are divided into different segmentation regions, thus completing the image segmentation task.

6. The method for generating a medical report on the severity of skin burns according to claim 5, characterized in that, The process of obtaining a medical report on the severity of skin burns based on statistical burn data from localized areas of varying burn degrees is as follows: Based on the proportion of the total burn area in the burn site and the proportion of local areas with different burn degrees in the total burn area, data on the proportion of the overall and local burn area are obtained. The average gray value of normal area pixels in the burned area is used as the skin color of the normal area. The greater the difference between the gray value of the skin color in different local burned areas and the gray value of the skin color in the normal area, the more severe the burn. The severity of the burn in the local area constitutes the severity of the burn in the overall burned area, thus obtaining the data on the severity of the burn in the local burned areas of the overall burn. Determine whether the burned area crosses a joint. If not, no additional annotation is needed in the image; if so, additional note is needed in the medical report. By combining the patient's personal information, burn site image data, locally segmented burn areas marked in the burn site image, and additional information such as severity data and whether it crosses multiple joints, a medical report on the severity of the patient's skin burns can be obtained, and a corresponding report can be generated.

7. A medical report generation system for the severity of skin burns, characterized in that, include: The preprocessing module acquires images of the patient's burn sites and performs preprocessing on these images. An edge detection module performs edge detection on the preprocessed image of the burn area to obtain the obvious edge texture of the burn area. based on The operator performs edge detection on the preprocessed burn area image to obtain obvious edge textures in the image. The edge textures include the edge textures of the limb edges and the edge textures of the burn area. The edge textures of the burn area include the inner edges and the outer edges. The appearance of the inner edges indicates that there are local burn areas of different severity in the whole burn area. The limb edges are directly detected by edge detection. The judgment module determines whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area. If so, the obtained overall edge pixels are fitted with a contour to obtain the overall edge of the burn area. Before determining whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area, the process further includes: constructing a maximum bounding rectangle for the edge texture of the burn area; based on the positional relationship between the coordinates of the edge texture pixels and the coordinates of the center point of the rectangle, and the gradient values ​​of the edge texture pixels, obtaining the overall edge coefficients of the pixels in the edge texture within the bounding rectangle, specifically: ； in, Represents the pixels in the texture of the inner edge of the rectangle. The overall edge coefficient; Indicates the coordinates of the center point of the rectangle; Indicates the first The coordinates of each edge texture pixel; Indicates the first The shortest distance between each edge texture pixel and its bounding rectangle; Indicates the first Gradient values ​​of each edge texture pixel; The determination of whether the pixels contained in the obvious edge texture of the burn area are pixels of the overall edge of the burn area is specifically as follows: based on the relationship between the overall edge coefficient and the set threshold, it is determined whether the pixels in the edge texture within the outer rectangle belong to the pixels of the overall edge. If so, the contour of all the obtained overall edge pixels is fitted to obtain the overall edge of the burn area. The segmentation module downsamples the burn area and segments the downsampled burn area to obtain regions with different degrees of burn. The density coefficients are constructed by iterating through the pixels in the downsampled image using the mean-shift algorithm. based on The algorithm analyzes the density coefficient histogram to identify local regions where density centers may exist. The acquired local area is regarded as the positioning area. The bandwidth variation coefficient of the positioning area is obtained by considering the total number of pixels in different positioning areas and the density coefficient in the positioning area. Based on the bandwidth variation coefficient of the positioning area, the bandwidth size of the positioning area is obtained; the initial bandwidth parameter size of the non-positioning area is set, and the mean drift algorithm is applied to the non-positioning area to obtain the bandwidth size of the non-positioning area. Based on the bandwidth parameters of the mean drift algorithm in different regions, the local area with different burn degrees is located. The acquisition module statistically analyzes burn data from local areas with different degrees of burns to obtain a medical report on the severity of skin burns.

Images