A machine learning-based wound assessment method, apparatus, device, and medium

Abstract

The application relates to a machine learning-based wound evaluation method, device, equipment and medium, belonging to the field of intelligent medical treatment. The method comprises the following steps: a user uploads a wound original image; wound annotation information is obtained based on the wound original image and is returned to a user end, the wound annotation information comprises a wound contour, a wound size and a wound recovery stage, the wound contour comprises a conventional wound contour and a beautified wound contour, the wound recovery stage comprises a coagulation period, an inflammation period, a repair period and a maturation period, the user end comprises a conventional user end and a pure user end, the conventional wound contour is returned to the conventional user end, and the beautified wound contour is returned to the pure user end; and wound care suggestions are provided to the user end according to the wound annotation information. The application realizes rapid and convenient identification of a wound, returns wound information according to different needs of users, and provides reasonable wound care suggestions.

Description

Technical Field

[0001] This invention belongs to the field of smart medical technology, specifically relating to a wound assessment method, device, equipment, and medium based on machine learning. Background Technology

[0002] Wound care typically requires one-on-one interaction between a professional nurse and the patient, which is labor-intensive and time-consuming for both parties. This can easily lead to delayed treatment and worsening of the wound, especially for chronic wounds requiring prolonged treatment. Chronic wounds have long healing times and are prone to further deterioration during the healing process, necessitating frequent observation of the patient's wound by professional nurses to develop further treatment plans based on wound feedback. This method demands significant travel time and effort from both the patient and medical staff.

[0003] Wound observation and analysis are fundamental to wound treatment and care. With the development of machine learning technology and the continuous upgrading of medical equipment, data-driven machine learning research is widely applied in the medical field, particularly in the analysis and processing of medical images. Machine learning technology can assist doctors in interpreting images, analyzing valuable information within them, and achieving intelligent analysis of medical images. However, current machine learning-based wound assessment methods suffer from incomplete wound feature extraction, and for individuals with poor psychological resilience, such as young children and the elderly, the constant exposure to graphic images of wounds during the assessment process can be detrimental to their physical and mental health. Summary of the Invention

[0004] To address the aforementioned problems in the prior art, this invention provides a wound assessment method, apparatus, device, and medium based on machine learning.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A machine learning-based wound assessment method, the implementation of which includes the following steps:

[0007] User uploads original image of the wound;

[0008] Based on the original wound image, wound annotation information is obtained and transmitted back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an enhanced wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is transmitted back to the regular user terminal, and the enhanced wound outline is transmitted back to the clean user terminal.

[0009] Based on the wound labeling information, wound care suggestions are provided to the user.

[0010] Preferably, the step of obtaining wound annotation information based on the original wound image and transmitting it back to the user terminal includes:

[0011] The contour of the conventional wound is obtained based on an improved contour recognition method;

[0012] An approximate beautified contour is obtained based on the conventional wound contour, the approximate beautified contour is filled with color, and the approximate beautified contour is used to replace the conventional wound contour to obtain the beautified wound contour.

[0013] The wound size is obtained based on the conventional wound contour;

[0014] The wound recovery stage is obtained based on an improved wound recognition model.

[0015] Preferably, obtaining the conventional wound contour based on the improved contour recognition method includes:

[0016] Uniformly collect contour sampling points in the original wound image and obtain the coordinates of the contour sampling points, and obtain the contour centroid based on the coordinates of the contour sampling points;

[0017] Obtain the centroid height value of the contour sampling point, where the centroid height value is the Euclidean distance between the contour sampling point and the contour centroid;

[0018] Calculate the centroid height increment based on the centroid height value;

[0019] Based on the centroid height increment, obtain the centroid height increment sequence of the i-th contour sampling point;

[0020] The centroid height increment matrix is ​​obtained based on the centroid height increment sequence.

[0021] The outline of the conventional wound is depicted based on the centroid height increment matrix.

[0022] Preferably, obtaining the approximate beautified contour based on the conventional wound contour includes:

[0023] Randomly select positive integer coefficients with values ​​in the open interval (1, N), and divide the centroid height increment sequence into S non-overlapping intervals [1, d], [d+1, 2d], ... based on the positive integer coefficients, where S = N / d, d is the positive integer coefficient, and calculate the mean centroid height increment within the interval;

[0024] The feature matrix of the i-th contour sampling point is obtained based on the mean centroid height increment within the interval.

[0025] The feature matrix of all the contour sampling points is obtained by organizing the feature matrices of all the contour sampling points;

[0026] A preset template image outline is obtained, and the template outline feature matrix of the template image outline is acquired.

[0027] Calculate the regular path distance between the conventional wound contour and the template image contour;

[0028] Flip the regular wound contour to obtain a flipped regular wound contour, and obtain the flipped regular path distance between the flipped regular wound contour and the template image contour;

[0029] The conventional shape complexity of the conventional wound contour and the template shape complexity of the template image contour are obtained.

[0030] A similarity coefficient is calculated based on the regularized path distance, the flipped regularized path distance, the conventional shape complexity, and the template shape complexity.

[0031] Traverse the template image contours, select the template image contour with the smallest similarity coefficient as the approximate beautified contour, and use the color-filled approximate beautified contour to cover the regular wound contour to obtain the beautified wound contour.

[0032] Preferably, the formula for calculating the similarity coefficient is as follows: Where Q is the similarity coefficient, G is the conventional shape complexity, and G Z For template shape complexity, d DTM (F B ,F Z ) represents the distance of the flipped normalized path, d DTM (F,F Z () represents the normalized path distance.

[0033] Preferably, the improved wound recognition model includes a backbone extraction network CSPDarknet, an enhanced feature extraction network FPN, and a classification regressor YOLOhead.

[0034] Preferably, the backbone extraction network CSPDarknet includes a Focus block and a Resblock body block. The Focus block is used to sample the pixel values ​​of the original wound image into four feature maps with three channels (RGB), forming a total of twelve feature maps. The Resblock body block is used to extract features from the feature maps using four max pooling kernels.

[0035] The enhanced feature extraction network (FPN) is used to further extract feature layers and transform feature layers of different sizes into feature layers of the same size. The layers are connected in the channel dimension to complete feature fusion and obtain a multi-scale feature map. The feature layers include bottom-level features and top-level features.

[0036] The YOLOhead classifier and regressor includes a classification part and a regression part. The classification part performs classification of the multi-scale feature map by performing two convolutions and then one convolution. The regression part uses two convolutions for feature extraction and finally performs regression prediction using two convolutions respectively to obtain the wound recovery stage.

[0037] A machine learning-based wound assessment device for performing the wound assessment method described above includes an image uploading module, an information annotation module, and a suggestion provision module.

[0038] The image upload module is used by users to upload original images of the wound;

[0039] The information annotation module is used to obtain wound annotation information based on the original wound image and send it back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an aesthetically pleasing wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is sent back to the regular user terminal, and the aesthetically pleasing wound outline is sent back to the clean user terminal.

[0040] The suggestion module is used to provide wound care suggestions to the user based on the wound labeling information.

[0041] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the wound assessment method described above.

[0042] A storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the wound assessment method described above.

[0043] The beneficial effects of this invention are as follows:

[0044] (1) Obtain wound annotation information through the original wound image and send it back to the user terminal to provide the user with reasonable wound care suggestions, save time and medical resources for both doctors and patients, and reduce the situation where patients cannot seek medical treatment in time due to various reasons, resulting in wound deterioration.

[0045] (2) Obtain an approximate beautified outline through the regular wound outline, fill the approximate beautified outline with color and replace the regular wound outline with the approximate beautified outline to obtain the beautified wound outline. Use images similar to wounds, such as animals and cartoons, to cover the bloody wound images, reduce the visual impact on users, and users can freely choose the regular user terminal or the clean user terminal, which has a high degree of freedom.

[0046] (3) Based on the centroid height increment matrix, the contour of a regular wound is drawn. The position of the wound contour point relative to the centroid and other wound contour points can be obtained to locate and draw the contour of a regular wound.

[0047] (4) The original wound image is first processed by CSPDarknet for feature extraction. Then, FPN uses the multi-layer feature maps extracted by CSPDarknet to enhance feature extraction, that is, to fuse the multi-layer feature maps of CSPDarknet to form multi-scale feature maps. Finally, YOLOhead performs classification and prediction to accurately determine the stage of wound development.

[0048] (5) The pixel values ​​of the original wound image are sampled into four feature maps by the Focus block. There are three channels of RGB, forming a total of twelve feature maps. The large number of channels results in high processing efficiency and accuracy. The feature layers are further extracted by FPN and the feature layers of different sizes are transformed into feature layers of the same size. The feature layers include bottom features and top features, instead of only extracting top features, which further improves the accuracy of the model. Attached Figure Description

[0049] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0050] Figure 1 This is a flowchart of a machine learning-based wound assessment method according to the present invention. Detailed Implementation

[0051] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0052] Working principle and usage process of this invention:

[0053] Please see Figure 1 A machine learning-based wound assessment method includes:

[0054] S1: User uploads original image of the wound;

[0055] S2: Obtain wound annotation information based on the original wound image and send it back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an enhanced wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is sent back to the regular user terminal, and the enhanced wound outline is sent back to the clean user terminal.

[0056] S3: Provide wound care suggestions to the user based on the wound labeling information.

[0057] In this embodiment, obtaining wound annotation information based on the original wound image can be implemented through the following steps:

[0058] S201: Obtain the contour of the conventional wound based on the improved contour recognition method;

[0059] S202: Obtain an approximate beautified contour based on the conventional wound contour, fill the approximate beautified contour with color and replace the conventional wound contour with the approximate beautified contour to obtain the beautified wound contour, cover the bloody wound image with images similar to wounds such as animals and cartoons to reduce the visual impact on users, and users can freely choose between the conventional user client or the clean user client, with a high degree of freedom.

[0060] S203: Obtain the wound size based on the conventional wound contour;

[0061] S204: Obtain the wound recovery stage based on an improved wound recognition model, wherein the improved wound recognition model includes a backbone extraction network CSPDarknet, an enhanced feature extraction network FPN, and a classification regressor YOLOhead.

[0062] In this embodiment, the contour of the conventional wound is obtained based on an improved contour recognition method, which can be implemented through the following steps:

[0063] S201-1: Uniformly acquire contour sampling points in the original wound image and obtain the coordinates of the contour sampling points. Obtain the contour centroid based on the contour sampling point coordinates. The expression for the contour centroid is: Where Q(x0,y0) is the centroid of the contour, and N is the number of contour sampling points, (x i ,y i () represents the coordinates of the i-th contour sampling point;

[0064] S201-2: Obtain the centroid height value of the contour sampling point, wherein the centroid height value is the Euclidean distance between the contour sampling point and the contour centroid;

[0065] S201-3: Calculate the centroid height increment based on the stated centroid height value. The calculation formula is h. i，j =g j -g i i, j = 1, 2, ..., N, where h i,j g represents the centroid height increment of the i-th contour sampling point and the j-th contour sampling point. i Let g be the centroid height value of the i-th contour sampling point. j Let the centroid height value be the j-th contour sampling point;

[0066] S201-4: Obtain the centroid height increment sequence of the i-th contour sampling point based on the centroid height increment, wherein the expression for the centroid height increment sequence is H. i =(h i，j h i，i+1 , ..., h i，N h i，1 , ..., h i，i-1 ) T , i = 1, 2, ..., N, where H i h is the centroid height increment sequence of the i-th contour sampling point. i,i h is the centroid height increment between the i-th contour sampling point and the i-th contour sampling point, with a value of 0. i,i+1 h represents the centroid height increment between the i-th contour sampling point and the (i+1)-th contour sampling point. i,N h represents the centroid height increment between the i-th contour sampling point and the N-th contour sampling point. i,1 h represents the centroid height increment between the i-th contour sampling point and the 1-th contour sampling point. i,i-1 This represents the centroid height increment of the i-th contour sampling point and the (i-1)-th contour sampling point;

[0067] S201-5: Obtain the centroid height increment matrix based on the centroid height increment sequence. The expression for the centroid height increment matrix is ​​L(M) = (H1, H2, ..., H...). N-1 H N ), where L(M) is the centroid height increment matrix, H1 is the centroid height increment sequence of the first contour sampling point, H2 is the centroid height increment sequence of the second contour sampling point, H N-1 H is the centroid height increment sequence of the (N-1)th contour sampling point. N This is the centroid height increment sequence of the Nth contour sampling point;

[0068] S201-6: Delineate the outline of the conventional wound based on the centroid height increment matrix.

[0069] In this embodiment, an approximate beautified contour is obtained based on the conventional wound contour, which can be implemented through the following steps:

[0070] S202-1: Randomly select positive integer coefficients with values ​​ranging from the open interval (1, N). Based on these positive integer coefficients, divide the centroid height increment sequence into S disjoint intervals [1, d], [d+1, 2d], ..., where S = N / d, and d is the positive integer coefficient. Calculate the mean centroid height increment within each interval using the following formula: t = 1, 2, ..., S, where c i,th is the mean centroid height increment within the t-th interval of the i-th contour sampling point. i,j This represents the centroid height increment of the i-th contour sampling point and the j-th contour sampling point;

[0071] S202-2: Obtain the feature matrix of the i-th contour sampling point based on the average centroid height increment within the interval, and the expression of the feature matrix is ​​C. i =(c i，1 c i，2 c i，S ) T , i = 1, 2, ..., N, where C i Let i be the feature matrix of the i-th contour sampling point;

[0072] S202-3: Organize the feature matrices of all the aforementioned contour sampling points to obtain the contour feature matrix, expressed as F = (C1, C2, ..., C...). N ), where F is the contour feature matrix;

[0073] S202-4: Preset template image outline, wherein the template image outline is the outline of an aesthetically pleasing template image such as an animal, plant, or cartoon; obtain the template outline feature matrix of the template image outline, wherein the expression of the template outline feature matrix is ​​F. Z =(C Z1 C Z2 C ZN ), where F Z C is the template outline feature matrix. Zl Let be the feature matrix of the l-th contour sampling point in the template image, where l = 1, 2, ..., N;

[0074] S202-5: Calculate the regular path distance between the conventional wound contour and the template image contour, using the formula d. DTM (F,F Z )=min∑|C i -C Zl | 2 , i, l = 1, 2, ..., N, where d DTM (F,F Z () represents the normalized path distance;

[0075] S202-6: Flip the conventional wound contour to obtain a flipped conventional wound contour, and obtain the flipped normalization path distance d between the flipped conventional wound contour and the template image contour. DTM (F B ,F Z );

[0076] S202-7: Obtain the conventional shape complexity of the conventional wound contour and the template shape complexity of the template image contour. The expression for the conventional shape complexity is: Where G is the complexity of the regular shape, and std is the standard deviation;

[0077] S202-8: Calculate the similarity coefficient based on the regularized path distance, the flipped regularized path distance, the conventional shape complexity, and the template shape complexity. The calculation formula is as follows: Where Q is the similarity coefficient, and G Z For template shape complexity;

[0078] S202-9: Traverse the template image contours, select the template image contour with the smallest similarity coefficient as the approximate beautified contour, and use the color-filled approximate beautified contour to cover the regular wound contour to obtain the beautified wound contour.

[0079] In this embodiment, the wound recovery stage is obtained based on an improved wound recognition model, which can be implemented through the following steps:

[0080] S204-1: The backbone extraction network CSPDarknet includes a Focus block and a Resblock body block. The Focus block is used to sample the pixel values ​​of the original wound image into four feature maps with three channels (RGB), forming a total of twelve feature maps. The large number of channels results in high processing efficiency and accuracy. The Resblock body block is used to extract features from the feature maps using four max pooling kernels.

[0081] S204-2: The enhanced feature extraction network FPN is used to further extract feature layers and transform feature layers of different sizes into feature layers of the same size. It is connected in the channel dimension to complete feature fusion and obtain a multi-scale feature map. The feature layer includes bottom-level features and top-level features, instead of only extracting top-level features, which further improves the model accuracy.

[0082] S204-3: The YOLOhead classifier includes a classification part and a regression part. The classification part performs classification of the multi-scale feature map by performing two convolutions and then one convolution. The regression part uses two convolutions for feature extraction and finally performs regression prediction using two convolutions respectively to obtain the wound recovery stage.

[0083] In this embodiment, wound care suggestions are provided to the user based on the wound marking information, which can be implemented through the following steps:

[0084] The wound care recommendations were obtained by crawling online resources. During the clotting phase, it is recommended to stop bleeding and disinfect promptly. If the wound is large, medical attention should be sought immediately. During the inflammatory phase, it is recommended to destroy bacteria and remove necrotic tissue to lay the foundation for tissue regeneration and repair. During the repair phase, it is recommended to prevent wound hyperplasia. During the maturation phase, it is recommended to improve the appearance and function of the injured area.

[0085] A machine learning-based wound assessment device includes an image uploading module, an information annotation module, and a suggestion provision module;

[0086] The image upload module is used by users to upload original images of the wound;

[0087] The information annotation module is used to obtain wound annotation information based on the original wound image and send it back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an aesthetically pleasing wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is sent back to the regular user terminal, and the aesthetically pleasing wound outline is sent back to the clean user terminal.

[0088] The suggestion module is used to provide wound care suggestions to the user based on the wound labeling information.

[0089] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0090] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0091] The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof. The computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages—such as Java, Smalltalk, and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A wound assessment method based on machine learning, characterized in that, The implementation of the wound assessment method includes the following steps: User uploads original image of the wound; Based on the original wound image, wound annotation information is obtained and transmitted back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an enhanced wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is transmitted back to the regular user terminal, and the enhanced wound outline is transmitted back to the clean user terminal. Based on the wound labeling information, wound care suggestions are provided to the user. The step of obtaining wound annotation information based on the original wound image and transmitting it back to the user terminal includes: The contour of the conventional wound is obtained based on an improved contour recognition method; An approximate beautified contour is obtained based on the conventional wound contour, the approximate beautified contour is filled with color, and the approximate beautified contour is used to replace the conventional wound contour to obtain the beautified wound contour. The wound size is obtained based on the conventional wound contour; The wound recovery stage is obtained based on an improved wound recognition model; The improved wound recognition model includes a backbone extraction network CSPDarknet, an enhanced feature extraction network FPN, and a classification and regressor YOLOhead. The backbone extraction network CSPDarknet includes a Focus block and a Resblock body block. The Focus block is used to sample the pixel values ​​of the original wound image into four feature maps with three channels (RGB), forming a total of twelve feature maps. The Resblock body block is used to extract features from the feature maps using four max pooling kernels. The enhanced feature extraction network (FPN) is used to further extract feature layers and transform feature layers of different sizes into feature layers of the same size. The layers are connected in the channel dimension to complete feature fusion and obtain a multi-scale feature map. The feature layers include bottom-level features and top-level features. The YOLOhead classifier and regressor includes a classification part and a regression part. The classification part performs classification of the multi-scale feature map by performing two convolutions and then one convolution. The regression part uses two convolutions for feature extraction and finally performs regression prediction using two convolutions respectively to obtain the wound recovery stage.

2. The wound assessment method according to claim 1, characterized in that, The step of obtaining the conventional wound contour based on the improved contour recognition method includes: Uniformly collect contour sampling points in the original wound image and obtain the coordinates of the contour sampling points, and obtain the contour centroid based on the coordinates of the contour sampling points; Obtain the centroid height value of the contour sampling point, where the centroid height value is the Euclidean distance between the contour sampling point and the contour centroid; Calculate the centroid height increment based on the centroid height value; Based on the centroid height increment, obtain the centroid height increment sequence of the i-th contour sampling point; The centroid height increment matrix is ​​obtained based on the centroid height increment sequence. The outline of the conventional wound is depicted based on the centroid height increment matrix.

3. The wound assessment method according to claim 2, characterized in that, The process of obtaining an approximate beautified contour based on the conventional wound contour includes: Randomly select positive integer coefficients with values ​​in the open interval (1, N), and divide the centroid height increment sequence into S non-overlapping intervals [1, d], [d+1, 2d], ... based on the positive integer coefficients, where S = N / d, d is the positive integer coefficient, and calculate the mean centroid height increment within the interval; The feature matrix of the i-th contour sampling point is obtained based on the mean centroid height increment within the interval. The feature matrix of all the contour sampling points is obtained by organizing the feature matrices of all the contour sampling points; A preset template image outline is obtained, and the template outline feature matrix of the template image outline is acquired. Calculate the regular path distance between the conventional wound contour and the template image contour. The formula for calculating the regular path distance is as follows: ; where d DTM (F,F Z ) represents the regularized path distance; F is the contour feature matrix, expressed as: ;F Z The template outline feature matrix is ​​expressed as follows: C Zl For the template image, the first l The feature matrix of each contour sampling point l =1,2,…,N;C i Let be the feature matrix of the i-th contour sampling point, expressed as: ; Flip the regular wound contour to obtain a flipped regular wound contour, and obtain the flipped regular path distance between the flipped regular wound contour and the template image contour; The conventional shape complexity of the conventional wound contour and the template shape complexity of the template image contour are obtained. A similarity coefficient is calculated based on the regularized path distance, the flipped regularized path distance, the conventional shape complexity, and the template shape complexity. Traverse the template image contours, select the template image contour with the smallest similarity coefficient as the approximate beautified contour, and use the color-filled approximate beautified contour to cover the regular wound contour to obtain the beautified wound contour.

4. The wound assessment method according to claim 3, characterized in that, The formula for calculating the similarity coefficient is as follows: Where Q is the similarity coefficient, G is the conventional shape complexity, and Gc is the shape complexity. Z For template shape complexity, d DTM (F B ,F Z ) represents the distance of the flipped normalized path, d DTM (F,F Z () represents the normalized path distance.

5. A wound assessment device based on machine learning, characterized in that, The device is applied to the wound assessment method as described in any one of claims 1-4, and includes an image uploading module, an information annotation module, and a suggestion provision module; The image upload module is used by users to upload original images of the wound; The information annotation module is used to obtain wound annotation information based on the original wound image and send it back to the user terminal. The wound annotation information includes wound outline, wound size, and wound recovery stage. The wound outline includes a regular wound outline and an aesthetically pleasing wound outline. The wound recovery stage includes the coagulation stage, inflammation stage, repair stage, and maturation stage. The user terminal includes a regular user terminal and a clean user terminal. The regular wound outline is sent back to the regular user terminal, and the aesthetically pleasing wound outline is sent back to the clean user terminal. The suggestion module is used to provide wound care suggestions to the user based on the wound labeling information.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the wound assessment method as described in any one of claims 1-4.

7. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the wound assessment method as described in any one of claims 1-4.

Images