Lymphedema identification method and device for postoperative breast cancer surgery

Abstract

The embodiments of the present disclosure disclose a method and device for identifying lymphedema after breast cancer surgery. A specific implementation of the method includes: constructing a three-dimensional upper limb model based on a sequence of collected upper limb images; performing upper limb status recognition based on the upper limb image sequence and the three-dimensional upper limb model to generate upper limb status information; generating lymphedema recognition information based on the upper limb status information, the collected chief complaint information, breast cancer treatment information corresponding to the subject to be identified, and a pre-trained lymphedema recognition model; in response to the lymphedema recognition information indicating that the subject to be identified has lymphedema, generating rehabilitation advice information corresponding to the subject to be identified based on the lymphedema recognition information. This implementation allows patients after breast cancer surgery to detect their own lymphedema as early as possible, so that when lymphedema occurs, patients can receive early prevention and treatment.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of computer technology, and more particularly to a method and apparatus for identifying lymphedema after breast cancer surgery. Background Art

[0002] Breast Cancer Related Lymphedema (BCRL) is a common chronic complication after breast cancer surgery. It is caused by obstruction of the return of protein-rich lymph fluid after breast cancer surgery, which is then retained in the interstitial space and causes edema, leading to circulatory disorders in the lymphatic system. Symptoms include limited range of motion, pain, fat deposition, fibrosis, and other changes in the upper limbs, and even severe skin infections. Because it is a chronic complication after surgery, early detection of lymphedema is crucial for its treatment.

[0003] The above information disclosed in this Background section is only for enhancement of understanding of the background of the inventive concept and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0004] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0005] Some embodiments of the present disclosure provide methods and devices for identifying lymphedema after breast cancer surgery to solve the technical problems mentioned in the above background technology section.

[0006] In a first aspect, some embodiments of the present disclosure provide a method for identifying lymphedema after breast cancer surgery, the method comprising: constructing a three-dimensional upper limb model based on a sequence of upper limb images obtained, wherein the three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement areas of the object to be identified and obtained from multiple angles; performing upper limb state recognition based on the upper limb image sequence and the three-dimensional upper limb model to generate upper limb state information, wherein the upper limb state information includes: upper limb dimensional features and upper limb skin features, wherein the upper limb dimensional features include: upper limb arm circumference, and the upper limb skin features include: Upper limb color, allergy characteristics, and scar characteristics; generating lymphedema identification information based on the upper limb status information, the collected chief complaint information, the breast cancer surgical treatment information corresponding to the above-mentioned subject to be identified, and a pre-trained lymphedema recognition model, wherein the above-mentioned chief complaint information is a description of the upper limb status provided by the above-mentioned subject to be identified, and the above-mentioned breast cancer surgical treatment information includes: surgical site, surgical incision type, and radiotherapy information; in response to the above-mentioned lymphedema identification information indicating that the above-mentioned subject to be identified has lymphedema, generating rehabilitation recommendation information corresponding to the above-mentioned subject to be identified based on the above-mentioned lymphedema identification information.

[0007] In a second aspect, some embodiments of the present disclosure provide a lymphedema identification device for use after breast cancer surgery, the device comprising: a construction unit configured to construct a three-dimensional upper limb model based on a sequence of upper limb images acquired, wherein the three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement areas acquired at multiple angles for an object to be identified; an upper limb state identification unit configured to perform upper limb state identification based on the upper limb image sequence and the three-dimensional upper limb model to generate upper limb state information, wherein the upper limb state information comprises upper limb dimensional features and upper limb skin features, wherein the upper limb dimensional features comprise upper limb arm circumference, and the upper limb skin features comprise: Upper limb color, allergy characteristics and scar characteristics; a first generation unit is configured to generate lymphedema identification information based on the above-mentioned upper limb status information, the collected chief complaint information, the breast cancer surgery treatment information corresponding to the above-mentioned object to be identified and a pre-trained lymphedema recognition model, wherein the above-mentioned chief complaint information is a description of the upper limb status provided by the above-mentioned object to be identified, and the above-mentioned breast cancer surgery treatment information includes: surgical site, surgical incision type and radiotherapy information; the second generation unit is configured to respond to the above-mentioned lymphedema identification information to characterize the presence of lymphedema in the above-mentioned object to be identified, and generate rehabilitation recommendation information corresponding to the above-mentioned object to be identified based on the above-mentioned lymphedema identification information.

[0008] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0009] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0010] The aforementioned embodiments of the present disclosure have the following beneficial effects: The lymphedema identification methods for breast cancer surgery, as applied to some embodiments of the present disclosure, can enable breast cancer patients to detect lymphedema as early as possible, allowing for early prevention and treatment of lymphedema. Specifically, because lymphedema is a common chronic complication after breast cancer surgery and can occur months to years after surgery, patients need to spontaneously determine whether they have lymphedema. Based on this, the lymphedema identification methods for breast cancer surgery, as applied to some embodiments of the present disclosure, first construct a three-dimensional upper limb model based on a sequence of acquired upper limb images. The three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement regions acquired from multiple angles for the subject to be identified. By combining the upper limb images to construct the local three-dimensional model (the upper limb three-dimensional model) corresponding to the at least three upper limb measurement regions, the upper limb corresponding to the subject to be identified can be effectively modeled. Next, upper limb status recognition is performed based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb status information. The upper limb status information includes upper limb dimensional features and upper limb skin features. The upper limb dimensional features include upper limb arm circumference, and the upper limb skin features include upper limb color, allergy characteristics, and scar characteristics. By combining the upper limb images and the upper limb three-dimensional model, the upper limb dimensional features and upper limb skin features of the subject to be identified are identified. Next, lymphedema identification information is generated based on the upper limb status information, the collected chief complaint information, the corresponding breast cancer surgical treatment information of the subject to be identified, and a pre-trained lymphedema identification model. The chief complaint information is a description of the upper limb status provided by the subject to be identified, and the breast cancer surgical treatment information includes surgical site, surgical incision type, and radiotherapy information. The pre-trained lymphedema identification model automatically combines the upper limb status information, chief complaint information, and breast cancer surgical treatment information to generate lymphedema identification information. In this way, patients can quickly identify whether they have lymphedema by collecting a sequence of upper limb images of their own upper limbs without going to the hospital. Finally, in response to the lymphedema identification information indicating that the above-mentioned object to be identified has lymphedema, rehabilitation advice information corresponding to the above-mentioned object to be identified is generated based on the lymphedema identification information. In this way, matching rehabilitation advice is generated based on the type of lymphedema, so that when the patient has milder symptoms, he or she can relieve lymphedema at home by combining rehabilitation advice. In this way, breast cancer patients after surgery can discover their own lymphedema as early as possible, so that patients can take early preventive measures when lymphedema occurs. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0012] Figure 1 is a flow chart of some embodiments of a method for identifying lymphedema after breast cancer surgery according to the present disclosure;

[0013] Figure 2 It is a schematic diagram of the location of the upper limb measurement area;

[0014] Figure 3 It is a comparison diagram of the upper limb image, the first channel image, the second channel image, and the third channel image;

[0015] Figure 4 It is a schematic diagram of the generation process of the image block set corresponding to the first channel image;

[0016] Figure 5 1 is a schematic structural diagram of some embodiments of a device for identifying lymphedema after breast cancer surgery according to the present disclosure;

[0017] Figure 6 It is a structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0018] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0019] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0020] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0024] refer to Figure 1 , shows a process 100 of some embodiments of the method for identifying lymphedema after breast cancer surgery according to the present disclosure. The method for identifying lymphedema after breast cancer surgery includes the following steps:

[0025] Step 101: construct an upper limb three-dimensional model based on the acquired upper limb image sequence.

[0026] In some embodiments, an entity (e.g., a computing device) executing a method for identifying lymphedema after breast cancer surgery can construct a three-dimensional upper limb model based on a sequence of captured upper limb images. The upper limb image sequence can be multiple images of the upper limb of a subject to be identified, captured using a mobile device with image acquisition capabilities (e.g., a mobile phone). Specifically, during upper limb image acquisition, the subject can have their arm rest horizontally on a flat surface in a relaxed position. The subject can flip their hand and rotate their upper limb to capture images of their upper limb from multiple angles. This eliminates the need for specialized equipment for image acquisition, thereby reducing the image acquisition complexity for the patient. The subject to be identified can be a patient undergoing breast cancer surgery for lymphedema identification. The three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement regions captured from multiple angles for the subject to be identified. In practice, the upper limb consists of six parts: shoulder, upper arm, elbow, forearm, wrist, and hand. Lymphedema primarily occurs in the upper arm, forearm, and hand. Therefore, setting at least three measurement regions can reduce the data processing required to construct the three-dimensional model. In particular, for the upper arm and forearm, local areas in the upper arm and forearm (for example, an area with a width of 5 cm in the middle of the upper arm and an area with a width of 5 cm in the middle of the forearm) can be used as upper limb measurement areas to further reduce the amount of data processing during three-dimensional modeling.

[0027] For example, see Figure 2 The schematic diagram of the location of the upper limb measurement area is shown in FIG. Figure 2 Three upper limb measurement regions are shown: the upper limb measurement region corresponding to the hand, the upper limb measurement region corresponding to a local area in the forearm, and the upper limb measurement region corresponding to a local area in the upper arm. By performing 3D modeling on only the upper limb measurement regions, the amount of data processing can be effectively reduced.

[0028] As another example, an upper limb image in an upper limb image sequence may contain depth information. Therefore, first, the execution entity may crop a local image corresponding to the upper limb measurement area from the upper limb image. Secondly, edge feature point matching is performed on the at least three local images obtained. Then, based on the edge feature point matching results, the relative position and orientation of the local image in the at least one local image obtained relative to the other local images is estimated using the RANSAC (RANdom SApmle Consensus) algorithm. Next, in combination with the depth information contained in the upper limb image, the pixel points in the local image are converted from two-dimensional image coordinates in the image coordinate system to three-dimensional coordinates in the geodetic coordinate system. Furthermore, the three-dimensional coordinates in the obtained three-dimensional coordinate set are subjected to coordinate system one (for example, by means of three-dimensional point cloud registration). Then, surface reconstruction and texture mapping are performed on the three-dimensional coordinate set after coordinate system one to obtain the above-mentioned local upper limb model.

[0029] It should be noted that the computing device described above can be either hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is software, it can be installed in the hardware devices listed above. It can be implemented as multiple software or software modules, for example, to provide distributed services, or as a single software or software module. No specific limitations are given here.

[0030] In some optional implementations of some embodiments, the execution subject constructs the upper limb three-dimensional model based on the acquired upper limb image sequence, including:

[0031] In the first step, for each upper limb image in the upper limb image sequence, the following first processing step is performed:

[0032] The first sub-step: dividing the upper limb image along the color channel to obtain a first channel image, a second channel image, and a third channel image.

[0033] The first channel image is an upper limb image under the blue channel, the second channel image is an upper limb image under the green channel, and the third channel image is an upper limb image under the red channel.

[0034] As an example, the code to segment an image along its color channels could be as follows:

[0035] import cv2

[0036] image = cv2.imread("storage path of upper limb image")

[0037] image_rgb = cv2.cvtColor(img,cv2.COLOR_BGR2RGB)

[0038] r_channel,g_channel,b_channel = cv2.split(image_rgb)

[0039] Among them, r_channel corresponds to the third channel image, g_channel corresponds to the second channel image, and b_channel corresponds to the first channel image.

[0040] As yet another example, see Figure 3 The following figure shows a comparison of an upper limb image, a first-channel image, a second-channel image, and a third-channel image. Analysis reveals that the upper limb-background boundary is more distinct in the third-channel image than in the first-channel image. This is due to the fact that, for example, the skin color of Asian individuals varies from yellow to red, pink to white. When the background color of an upper limb image is similar to the skin color of the upper limb, the boundary between the upper limb and the background in the third-channel image is unclear, hindering the subsequent extraction of key points along the upper limb boundary. Furthermore, based on the color distribution on the color wheel, green is close to yellow. Therefore, when the background color of an upper limb image is similar to the skin color of the upper limb, the boundary between the upper limb and the background in the second-channel image is more blurred than in the first-channel image and more distinct than in the third-channel image. Therefore, the first-channel image is used as the baseline. Furthermore, single-channel image processing can effectively reduce data processing compared to processing upper limb images in RGB (Red-Green-Blue) format.

[0041] Second sub-step: Divide the first channel image into a set of image blocks.

[0042] The image blocks in the above image block set have the same image block size. In practice, the first channel image can be divided into the image block set by uniform division.

[0043] The third sub-step: for each image block in the above image block set, perform the following second processing step:

[0044] Step 1: Perform image block type recognition on the above image block to generate an image block type.

[0045] The image block type represents the part type of the upper limb part corresponding to the image block.

[0046] In practice, a lightweight classification model can be built to identify the image block type. This model can be constructed using five serially connected convolutional layers and one fully connected layer to classify the image block type. The classification model can be trained using supervised training, using image blocks labeled with their type as training samples and sample labels.

[0047] For example, see Figure 4 The schematic diagram of the generation process of the image block set corresponding to the first channel image is shown, wherein the above-mentioned execution entity can divide the first channel image into 2×5 image blocks, a total of 10 image blocks, as the image block set. The relative positions of the 10 image blocks are fixed. The advantage of block processing is that when the key point corresponding to the first key point information is not extracted in the image block corresponding to the first channel image, the key point extraction of the image blocks corresponding to the second and third channel images can be skipped, thereby reducing the amount of data processing.

[0048] Step 2: In response to the image block type being consistent with the part type corresponding to any of the at least three upper limb measurement areas, key point recognition is performed on the image block to generate first key point information corresponding to the image block.

[0049] The first key point information includes: key point coordinates, image block index, and key point feature vector. In practice, the first key point information can be obtained by edge detection, such as performing edge detection on the image block using the Canny operator. The key point coordinates represent the image coordinates of the edge key point corresponding to the first key point information. The key point feature vector represents the feature vector of the edge key point corresponding to the first key point information, such as using a 256-dimensional feature vector for representation. The image block index represents the position of the above-mentioned image block in the image block set.

[0050] Step 3: Based on the second channel image and the third channel image, perform key point information filtering on the obtained first key point information group to obtain a key point information group.

[0051] Optionally, the key point information group obtained by filtering the first key point information group according to the second channel image and the third channel image to obtain the key point information group includes:

[0052] For each first key point information in the first key point information group, the following third processing step is performed:

[0053] S1: performing key point recognition on an image block in the second channel image corresponding to the image block index included in the first key point information to obtain second key point information.

[0054] The second key point information includes: key point coordinates and key point feature vectors.

[0055] In practice, the second key point information can be obtained by edge detection, such as using the Canny operator, on the image blocks in the second channel image corresponding to the image block index included in the first key point information. Since the first channel image, the second channel image, and the third channel image differ only in the color channel, the image block index can be used to locate the image blocks in the different channel images for further key point identification without processing the entire image, greatly reducing the amount of data processing.

[0056] S2: In response to the fact that the coordinate distance between the key point coordinates included in the first key point information and the key point coordinates included in the second key point information is less than the preset coordinate distance, determine the feature similarity between the key point feature vector included in the first key point information and the key point feature vector included in the second key point information to obtain the first feature similarity.

[0057] In practice, since key point feature vectors are high-dimensional feature vectors, directly calculating the similarity between vectors requires a large amount of data processing. In particular, when there are multiple key points in an image block, a greater number of similarities need to be calculated. Therefore, first comparing the distances between coordinates before calculating the similarity can effectively reduce the amount of data processing. Specifically, the cosine similarity can be calculated to determine the feature similarity between the key point feature vector included in the first key point information and the key point feature vector included in the second key point information, thereby obtaining the first feature similarity.

[0058] S3: In response to the first feature similarity being greater than or equal to a preset feature similarity, determining the first key point information as key point information in the key point information group.

[0059] S4: In response to the first feature similarity being less than a preset feature similarity, performing key point recognition on the image block in the third channel image corresponding to the image block index included in the first key point information to obtain third key point information.

[0060] The third key point information includes: key point coordinates and key point feature vectors.

[0061] In practice, the key point recognition method in "S2" can be used to perform key point recognition on the image block in the third channel image that corresponds to the image block index included in the first key point information to obtain the third key point information. No further details will be given here. Specifically, since the boundaries of the first channel image and the second channel image are clearer than those of the third channel image, the vast majority of matching key points can be determined by calculating the similarity of the feature points between the second channel image and the third channel image. For those that do not match, the third channel image is combined for supplementation. Compared with the method of calculating the similarity twice for the first key point information corresponding to the key point, and the second key point information corresponding to the key point, and the third key point information corresponding to the key point, when the key point corresponding to the first key point information matches the key point corresponding to the second key point information, one feature matching can be reduced, thereby further reducing the amount of data processing.

[0062] S5: In response to the fact that the coordinate distance between the key point coordinates included in the above-mentioned first key point information and the key point coordinates included in the above-mentioned third key point information is less than the preset coordinate distance, determine the feature similarity between the key point feature vector included in the above-mentioned first key point information and the key point feature vector included in the above-mentioned third key point information to obtain a second feature similarity.

[0063] Specifically, the cosine similarity may be calculated to determine the feature similarity between the key point feature vector included in the first key point information and the key point feature vector included in the third key point information, thereby obtaining the second feature similarity.

[0064] S6: In response to the second feature similarity being greater than or equal to a preset feature similarity, determining the first key point information as key point information in the key point information group.

[0065] Step 4: Construct a three-dimensional model of the upper limb based on the obtained set of key point information groups to obtain the above-mentioned three-dimensional model of the upper limb.

[0066] In practice, a key point information group in a key point information group set corresponds to an upper limb image. First, the matching key points in the upper limb images collected at different angles can be determined by feature point matching. Since the key point information includes key point feature vectors, the matching key points in different upper limb images can be determined by vector similarity calculation. Secondly, based on the multiple pairs of matched key point information, the essential matrix is estimated. Specifically, the basic matrix can be estimated by combining multiple pairs of matched key point information, and the basic matrix can be converted into the essential matrix by combining the intrinsic parameters of the camera. Then, based on the essential matrix, the rotation vector and translation matrix are determined. Furthermore, the key point coordinates corresponding to the matched key points are mapped to three-dimensional space as point cloud data by a triangulation method in combination with the rotation vector and translation matrix. Finally, a three-dimensional model is reconstructed based on the point cloud data to obtain a three-dimensional model of the upper limb.

[0067] Step 102 : performing upper limb status recognition based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb status information.

[0068] In some embodiments, the execution entity may perform upper limb status recognition based on an upper limb image sequence and a three-dimensional upper limb model to generate upper limb status information. The upper limb status information includes upper limb dimensional features and upper limb skin features. The upper limb dimensional features include upper limb arm circumference. The upper limb skin features include upper limb color, allergy features, and scar features. The allergy feature characterizes the type of allergy on the upper limb skin. The scar feature characterizes the type of scar on the upper limb skin.

[0069] As an example, the above-mentioned execution subject can determine the above-mentioned upper limb dimensional characteristics in combination with the upper limb three-dimensional model. Specifically, since the upper limb three-dimensional model is a local three-dimensional model corresponding to at least three upper limb measurement areas, the upper limb arm circumference can be obtained by combining the model perimeter of the upper limb three-dimensional model in the upper limb measurement area. In particular, since there is a proportional relationship between the upper limb three-dimensional model and the actual upper limb size during modeling, it is necessary to map the upper limb arm circumference obtained by the upper limb three-dimensional model in combination with the upper limb size of the object to be tested when normal, as the upper limb arm circumference. Since it contains at least three measurement areas, at least three upper limb arm circumferences can be obtained.

[0070] As another example, the execution entity can combine a sequence of upper limb images and obtain the upper limb skin features through image recognition. Specifically, a recurrent neural network model can be used to perform image recognition on the upper limb images to obtain the upper limb skin features. In particular, since upper limb skin features include upper limb color, allergy characteristics, and scar characteristics, three classifiers need to be connected to the recurrent neural network model to output upper limb color, allergy characteristics, and scar characteristics.

[0071] In some optional implementations of some embodiments, the execution subject performs upper limb state recognition based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb state information, including:

[0072] In the first step, for each upper limb image in the upper limb image sequence, the following background segmentation steps are performed:

[0073] The first sub-step is to construct a segmentation region based on the key point information group corresponding to the upper limb image.

[0074] In practice, since the key points corresponding to the key point information group are edge points of the upper limbs, the closed area surrounded by the key points corresponding to the key point information group can be used as the segmentation area.

[0075] In the second sub-step, the background of the upper limb image is stripped according to the segmented area to obtain an upper limb image after background stripping.

[0076] In practice, the execution entity can set the pixel values outside the segmented area to 0 to obtain a background-stripped upper limb image. This prevents non-upper limb images from participating in subsequent image processing, reducing data processing volume. Furthermore, compared to segmentation, this approach ensures that the resulting background-stripped upper limb image has a consistent image size.

[0077] In the third sub-step, image feature extraction is performed on the upper limb image after background stripping using an image feature extraction network included in a pre-trained upper limb skin feature extraction model to generate an image feature map.

[0078] In practice, the image feature extraction network can use the FPN (Feature Pyramid Network) network as the backbone network to extract image features from upper limb images after background stripping from different receptive fields.

[0079] The fourth sub-step is to generate candidate upper limb colors by using the upper limb color classifier included in the upper limb skin feature extraction model and the image feature map.

[0080] The upper limb color classifier is a multi-classifier that adjusts the feature dimensions of the image feature map and ultimately outputs a 1×N feature vector. N represents the color categories that the upper limb color classifier can classify. In practice, to avoid the difficulty of collecting training samples for skin color refinement, and because skin colors are primarily yellow, red, pink, white, and black, N can be set to 6. "6" corresponds to yellow, red, pink, white, black, and other colors.

[0081] The fifth sub-step is to generate candidate allergy features and candidate scar features through the allergy and scar location recognition network included in the upper limb skin feature extraction model and the above-mentioned image feature map.

[0082] In practice, in order to improve the recognition speed, the allergy and scar location recognition network can use ResNet-50 as the backbone network and connect two locators to generate candidate allergy features and candidate scar features respectively.

[0083] In the second step, the obtained candidate upper limb color set, candidate allergy feature set and candidate scar feature set are respectively voted and screened to obtain the upper limb color included in the above upper limb status information.

[0084] The third step is to determine the upper limb arm circumference included in the upper limb status information based on the regional cross-sectional perimeter of the upper limb three-dimensional model in the at least three upper limb measurement areas.

[0085] In practice, since there are at least three upper limb measurement areas, at least three upper limb arm circumferences can be obtained.

[0086] Step 103 : Generate lymphedema identification information based on the upper limb status information, the collected chief complaint information, the breast cancer treatment information corresponding to the object to be identified, and the pre-trained lymphedema identification model.

[0087] In some embodiments, the execution entity can generate lymphedema identification information based on upper limb status information, collected chief complaint information, breast cancer surgical treatment information corresponding to the subject to be identified, and a pre-trained lymphedema recognition model. The chief complaint information is a description of the upper limb status provided by the subject to be identified. Breast cancer surgical treatment information includes surgical site, surgical incision type, and radiotherapy information. In practice, since the subject to be identified is a patient who has undergone breast cancer surgery, corresponding breast cancer treatment records are used as post-operative breast cancer treatment information. Specifically, surgical sites include: lower inner chest, upper inner chest, areola, lower outer chest, and upper outer chest. Surgical incisions include: transverse incisions, longitudinal incisions, and oblique incisions. Radiotherapy information includes: radiation markers and radiation locations. The radiation marker indicates whether the subject to be identified has undergone radiation therapy. The radiation location indicates the location of the radiation therapy. Radiation locations include: breast, chest wall, and lymph node regions. The lymphedema recognition model can use a pre-trained model (e.g., a GPT model) as the backbone network to generate lymphedema identification information.

[0088] Optionally, lymphedema identification information includes: edema severity, edema type, and edema confidence. In practice, edema severity represents the severity of the lymphedema of the subject to be identified. For example, edema severity may include: Class A, Class B, and Class C. Class A, Class B, and Class C represent different levels of lymphedema severity. Class A corresponds to the highest severity. Class C corresponds to the lowest severity. Edema type represents the location where lymphedema occurs. Edema confidence represents the combined confidence of edema severity and edema type. Specifically, edema confidence is obtained by weighted summation of the confidence corresponding to edema severity and the confidence corresponding to edema type.

[0089] Optionally, the lymphedema recognition model includes a chief complaint information encoder, a chief complaint information decoder, an input layer, K hidden layers, and an output layer. The input layer includes 8 neurons, and the output layer includes 2 neurons. The lymphedema recognition model and the upper limb skin feature extraction model are trained as a whole through supervised training. In practice, the chief complaint information encoder and decoder utilize an encoder and decoder based on the Transformer architecture. The structures of the chief complaint information encoder and decoder are symmetrical. The input layer takes as input the output of the chief complaint information decoder, upper limb arm circumference, upper limb color, allergy characteristics, scar characteristics, surgical site, surgical incision type, and radiotherapy information. 8 ≥ K ≥ 5. This avoids the problem of excessive computational overhead caused by deep hidden layers. The output layer includes two neurons that output edema severity and edema type, respectively. The edema severity and edema type each have a corresponding confidence level. The edema confidence level is calculated by weighted summing the confidence level corresponding to the edema severity and the confidence level corresponding to the edema type.

[0090] In some optional implementations of some embodiments, the execution entity generates lymphedema identification information based on the upper limb status information, the collected chief complaint information, the breast cancer treatment information corresponding to the subject to be identified, and a pre-trained lymphedema identification model, including:

[0091] The first step is to perform text encoding on the chief complaint information through the chief complaint information encoder to obtain the chief complaint information features.

[0092] In the second step, the chief complaint information features are decoded by the chief complaint information decoder to obtain the chief complaint description features.

[0093] In the third step, based on the chief complaint description features, the upper limb status information, and the breast cancer surgery treatment information, lymphedema identification information is generated through the input layer, the K hidden layers, and the output layer.

[0094] Step 104 : In response to the lymphedema identification information indicating that the subject to be identified has lymphedema, rehabilitation advice information corresponding to the subject to be identified is generated according to the lymphedema identification information.

[0095] In some embodiments, the execution subject may generate rehabilitation advice information corresponding to the subject to be identified in response to the lymphedema identification information indicating that the subject to be identified has lymphedema. The rehabilitation advice information is rehabilitation advice for different degrees of severity of lymphedema. In practice, for mild lymphedema, self-training and rehabilitation can be used to alleviate lymphedema if it is discovered in time. For severe lymphedema, medical treatment can be used to treat lymphedema.

[0096] In some optional implementations of some embodiments, the execution entity generates rehabilitation advice information corresponding to the subject to be identified based on the lymphedema identification information in response to the lymphedema identification information indicating that the subject to be identified has lymphedema, including:

[0097] In response to the edema confidence level being greater than a preset edema confidence level, the rehabilitation advice information is determined based on the edema severity and edema type included in the lymphedema identification information, using a pre-constructed advice information decision tree. In practice, an advice information decision tree can be constructed based on different edema severities and edema types, as well as the corresponding rehabilitation advice information. This allows for rapid location of rehabilitation advice information. For example, for severe lymphedema, medical treatment and surgery can be used, such as connecting lymphatic vessels to veins to allow lymph fluid to flow back through the veins to alleviate lymphedema. Alternatively, lymph node suppression can be used to stimulate the construction of new lymphatic pathways to alleviate lymphedema. Excess tissue removal or liposuction can also be used to alleviate upper limb swelling caused by lymphedema. For mild lymphedema, avoiding hot baths or excessive heating can be used to prevent lymph fluid accumulation. It is also recommended to avoid prolonged sitting to prevent lymph fluid accumulation. In addition, wearing a lymphedema compression sleeve can be recommended to control lymphedema.

[0098] In some optional implementations of some embodiments, the above method further includes:

[0099] The first step is to generate rehabilitation training information corresponding to the above-mentioned object to be identified based on the above-mentioned rehabilitation suggestion information and the object portrait corresponding to the above-mentioned object to be identified.

[0100] The above-mentioned rehabilitation training information includes: training movement type, movement frequency, and rehabilitation duration. In practice, different rehabilitation movements can be pre-set for different rehabilitation recommendations to relieve lymphedema. At the same time, different subjects to be identified have different basic information such as age. Therefore, the movement frequency, rehabilitation duration, and training movement type can be adjusted based on the subject's portrait to ensure that the rehabilitation training information is consistent with the physical condition of the subject to be identified.

[0101] In the second step, in response to the object to be identified starting a rehabilitation training task, real-time action information corresponding to the object to be identified is collected.

[0102] In practice, the real-time action information corresponding to the object to be identified can be collected in real time by a mobile device (e.g., a mobile phone). For example, the real-time action information can be obtained by pointing the mobile device toward the object to be identified to collect and identify the action execution status of the object to be identified in real time.

[0103] The third step is to determine the action matching degree based on the above rehabilitation training information and the above real-time action information.

[0104] In practice, the above-mentioned execution subject can extract the action type corresponding to the real-time action information by extracting key points, and count the execution frequency and duration of the action of the object to be identified, and compare it with the training action type, action frequency, and rehabilitation duration included in the rehabilitation training information to determine the action matching degree.

[0105] In the fourth step, in response to the action matching degree being less than the preset matching degree, an action correction reminder is initiated to the above-mentioned object to be identified.

[0106] In practice, the above-mentioned execution subject can initiate action correction reminders to the object to be identified through a mobile device to ensure that the object to be identified performs the training actions corresponding to the rehabilitation training information with a high degree of matching, thereby achieving better rehabilitation effects.

[0107] The aforementioned embodiments of the present disclosure have the following beneficial effects: The lymphedema identification methods for breast cancer surgery, as applied to some embodiments of the present disclosure, can enable breast cancer patients to detect lymphedema as early as possible, allowing for early prevention and treatment of lymphedema. Specifically, because lymphedema is a common chronic complication after breast cancer surgery and can occur months to years after surgery, patients need to spontaneously determine whether they have lymphedema. Based on this, the lymphedema identification methods for breast cancer surgery, as applied to some embodiments of the present disclosure, first construct a three-dimensional upper limb model based on a sequence of acquired upper limb images. The three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement regions acquired from multiple angles for the subject to be identified. By combining the upper limb images to construct the local three-dimensional model (the upper limb three-dimensional model) corresponding to the at least three upper limb measurement regions, the upper limb corresponding to the subject to be identified can be effectively modeled. Next, upper limb status recognition is performed based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb status information. The upper limb status information includes upper limb dimensional features and upper limb skin features. The upper limb dimensional features include upper limb arm circumference, and the upper limb skin features include upper limb color, allergy characteristics, and scar characteristics. By combining the upper limb images and the upper limb three-dimensional model, the upper limb dimensional features and upper limb skin features of the subject to be identified are identified. Next, lymphedema identification information is generated based on the upper limb status information, the collected chief complaint information, the corresponding breast cancer surgical treatment information of the subject to be identified, and a pre-trained lymphedema identification model. The chief complaint information is a description of the upper limb status provided by the subject to be identified, and the breast cancer surgical treatment information includes surgical site, surgical incision type, and radiotherapy information. The pre-trained lymphedema identification model automatically combines the upper limb status information, chief complaint information, and breast cancer surgical treatment information to generate lymphedema identification information. In this way, patients can quickly identify whether they have lymphedema by collecting a sequence of upper limb images of their own upper limbs without going to the hospital. Finally, in response to the lymphedema identification information indicating that the above-mentioned object to be identified has lymphedema, rehabilitation advice information corresponding to the above-mentioned object to be identified is generated based on the lymphedema identification information. In this way, matching rehabilitation advice is generated based on the type of lymphedema, so that when the patient has milder symptoms, he or she can relieve lymphedema at home by combining rehabilitation advice. In this way, breast cancer patients after surgery can discover their own lymphedema as early as possible, so that patients can take early preventive measures when lymphedema occurs.

[0108] Further references Figure 5 As an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of a lymphedema identification device for breast cancer surgery. These device embodiments are similar to Figure 1Corresponding to the method embodiments shown, the lymphedema identification device applied to breast cancer surgery can be specifically applied to various electronic devices.

[0109] like Figure 5 As shown, some embodiments of the lymphedema recognition device 500 applied to breast cancer surgery include: a construction unit 501, an upper limb state recognition unit 502, a first generation unit 503, and a second generation unit 504. The construction unit 501 is configured to construct an upper limb three-dimensional model based on the acquired upper limb image sequence, wherein the upper limb three-dimensional model is a local three-dimensional model corresponding to at least three upper limb measurement areas acquired at multiple angles for the object to be identified; the upper limb state recognition unit 502 is configured to perform upper limb state recognition based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb state information, wherein the upper limb state information includes: upper limb dimensional features, upper limb skin features, the upper limb dimensional features include: upper limb arm circumference, the upper limb skin features include: upper limb color, allergy features and scar features; the first generation unit Element 503 is configured to generate lymphedema identification information based on the upper limb status information, the collected chief complaint information, the breast cancer surgical treatment information corresponding to the above-mentioned object to be identified, and the pre-trained lymphedema identification model, wherein the chief complaint information is a description of the upper limb status provided by the above-mentioned object to be identified, and the above-mentioned breast cancer surgical treatment information includes: surgical site, surgical incision type and radiotherapy information; the second generation unit 504 is configured to respond to the above-mentioned lymphedema identification information to characterize the presence of lymphedema in the above-mentioned object to be identified, and generate rehabilitation advice information corresponding to the above-mentioned object to be identified based on the above-mentioned lymphedema identification information. It can be understood that the various units recorded in the lymphedema identification device 500 for post-breast cancer surgery are the same as those in the reference 500. Figure 1 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the lymphedema identification device 500 used in breast cancer surgery and the units included therein, and will not be described in detail here.

[0110] Reference below Figure 6 , which shows a structural schematic diagram of an electronic device (eg, a computing device) suitable for implementing some embodiments of the present disclosure. Figure 6 The electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure. Figure 6As shown, the computer device includes a processor, a memory and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory. The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, which, when executed, can enable the processor to execute any front-end page monitoring method. The processor is used to provide computing and control capabilities to support the operation of the entire computer device. The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium, which, when executed by the processor, can enable the processor to execute any front-end page monitoring method. The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 6 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present disclosure, and does not constitute a limitation on the computer device to which the solution of the present disclosure is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0111] It should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0112] In one embodiment, the processor is configured to execute a computer program stored in a memory to implement the following steps: constructing a three-dimensional upper limb model based on a sequence of upper limb images acquired, wherein the three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement areas acquired at multiple angles for the object to be identified; performing upper limb state recognition based on the sequence of upper limb images and the three-dimensional upper limb model to generate upper limb state information, wherein the upper limb state information includes upper limb dimensional features and upper limb skin features, wherein the upper limb dimensional features include upper limb arm circumference, and the upper limb skin features include: Upper limb color, allergy characteristics, and scar characteristics; generating lymphedema identification information based on the upper limb status information, the collected chief complaint information, the breast cancer surgical treatment information corresponding to the above-mentioned subject to be identified, and a pre-trained lymphedema recognition model, wherein the above-mentioned chief complaint information is a description of the upper limb status provided by the above-mentioned subject to be identified, and the above-mentioned breast cancer surgical treatment information includes: surgical site, surgical incision type, and radiotherapy information; in response to the above-mentioned lymphedema identification information indicating that the above-mentioned subject to be identified has lymphedema, generating rehabilitation recommendation information corresponding to the above-mentioned subject to be identified based on the above-mentioned lymphedema identification information.

[0113] An embodiment of the present disclosure also provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to the various embodiments of the method for identifying lymphedema after breast cancer surgery applied to the present disclosure.

[0114] The computer-readable storage medium may be an internal storage unit of the computer device described in the aforementioned embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a flash memory card, etc., provided on the computer device.

[0115] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0116] The above descriptions are merely some preferred embodiments of the present disclosure and illustrate the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present disclosure is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A method for identifying lymphedema after breast cancer surgery, characterized in that: include: Constructing a three-dimensional upper limb model based on the acquired upper limb image sequence, wherein the three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement areas acquired from multiple angles for the object to be identified; Performing upper limb status recognition based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb status information, wherein the upper limb status information includes upper limb dimensional features and upper limb skin features, the upper limb dimensional features include upper limb arm circumference, and the upper limb skin features include upper limb color, allergy features, and scar features; Lymphedema identification information is generated based on the upper limb status information, the collected chief complaint information, the breast cancer surgical treatment information corresponding to the subject to be identified, and a pre-trained lymphedema identification model, wherein the chief complaint information is a description of the upper limb status provided by the subject to be identified, and the breast cancer surgical treatment information includes: surgical site, surgical incision type, and radiotherapy information. The lymphedema identification model includes: a chief complaint information encoder, a chief complaint information decoder, an input layer, K hidden layers, and an output layer, wherein the input layer includes: 8 neurons, and the output layer includes: 2 neurons. The lymphedema identification model and the upper limb skin feature extraction model are trained as a whole through a supervised training method, wherein the upper limb skin feature extraction model is used to extract upper limb skin features; In response to the lymphedema identification information indicating that the subject to be identified has lymphedema, generating rehabilitation advice information corresponding to the subject to be identified based on the lymphedema identification information, The lymphedema identification information is generated based on the upper limb status information, the collected chief complaint information, the breast cancer treatment information corresponding to the subject to be identified, and the pre-trained lymphedema identification model, including: Performing text encoding on the chief complaint information by the chief complaint information encoder to obtain chief complaint information features; Decoding the features of the chief complaint information by the chief complaint information decoder to obtain a chief complaint description feature; According to the chief complaint description features, the upper limb status information and the breast cancer surgery treatment information, lymphedema identification information is generated through the input layer, the K hidden layers and the output layer.

2. The method according to claim 1, characterized in that The lymphedema identification information includes: edema severity, edema type, and edema confidence; and In response to the lymphedema identification information indicating that the subject to be identified has lymphedema, generating rehabilitation advice information corresponding to the subject to be identified based on the lymphedema identification information, including: In response to the edema confidence being greater than a preset edema confidence, the rehabilitation advice information is determined according to the edema severity and edema type included in the lymphedema identification information through a pre-constructed advice information decision tree.

3. The method according to claim 2, characterized in that The method further comprises: Generate rehabilitation training information corresponding to the subject to be identified based on the rehabilitation suggestion information and the subject portrait corresponding to the subject to be identified, wherein the rehabilitation training information includes: training action type, action frequency, and rehabilitation duration; In response to the subject to be identified starting a rehabilitation training task, collecting real-time motion information corresponding to the subject to be identified; determining a degree of motion matching based on the rehabilitation training information and the real-time motion information; In response to the action matching degree being less than a preset matching degree, an action correction reminder is initiated to the object to be identified.

4. The method according to claim 3, characterized in that The constructing of the upper limb three-dimensional model according to the acquired upper limb image sequence includes: For each upper limb image in the upper limb image sequence, the following first processing step is performed: Dividing the upper limb image along the color channel to obtain a first channel image, a second channel image, and a third channel image, wherein the first channel image is the upper limb image under the blue channel, the second channel image is the upper limb image under the green channel, and the third channel image is the upper limb image under the red channel; Dividing the first channel image into an image block set, wherein the image blocks in the image block set have the same image block size; For each image block in the set of image blocks, performing the following second processing step: Performing image block type recognition on the image block to generate an image block type, wherein the image block type represents a part type of an upper limb part corresponding to the image block; In response to the image block type being consistent with the part type corresponding to any of the at least three upper limb measurement areas, performing key point recognition on the image block to generate first key point information corresponding to the image block, wherein the first key point information includes: key point coordinates, image block index, and key point feature vector; performing key point information filtering on the obtained first key point information group according to the second channel image and the third channel image to obtain a key point information group; The three-dimensional model of the upper limb is constructed according to the obtained key point information group set to obtain the three-dimensional model of the upper limb.

5. The method according to claim 4, characterized in that The step of filtering the obtained first key point information group according to the second channel image and the third channel image to obtain the key point information group includes: For each first key point information in the first key point information group, perform the following third processing step: Performing key point recognition on an image block in the second channel image that corresponds to the image block index included in the first key point information to obtain second key point information, wherein the second key point information includes: key point coordinates and key point feature vectors; In response to a coordinate distance between a key point coordinate included in the first key point information and a key point coordinate included in the second key point information being less than a preset coordinate distance, determining a feature similarity between a key point feature vector included in the first key point information and a key point feature vector included in the second key point information to obtain a first feature similarity; In response to the first feature similarity being greater than or equal to a preset feature similarity, determining the first key point information as key point information in the key point information group; In response to the first feature similarity being less than a preset feature similarity, performing key point recognition on an image block in the third channel image that corresponds to the image block index included in the first key point information to obtain third key point information, wherein the third key point information includes: key point coordinates and a key point feature vector; In response to a coordinate distance between the key point coordinates included in the first key point information and the key point coordinates included in the third key point information being less than a preset coordinate distance, determining a feature similarity between a key point feature vector included in the first key point information and a key point feature vector included in the third key point information to obtain a second feature similarity; In response to the second feature similarity being greater than or equal to a preset feature similarity, the first key point information is determined as key point information in the key point information group.

6. The method according to claim 5, characterized in that The performing upper limb state recognition according to the upper limb image sequence and the upper limb three-dimensional model to generate upper limb state information includes: For each upper limb image in the upper limb image sequence, perform the following background segmentation steps: Constructing a segmentation region according to the key point information group corresponding to the upper limb image; performing image background stripping on the upper limb image according to the segmented region to obtain an upper limb image after background stripping; performing image feature extraction on the upper limb image after background stripping using an image feature extraction network included in a pre-trained upper limb skin feature extraction model to generate an image feature map; Generate candidate upper limb colors using the upper limb color classifier included in the upper limb skin feature extraction model and the image feature map; Generate candidate allergy features and candidate scar features through the allergy and scar location recognition network included in the upper limb skin feature extraction model and the image feature map; Voting and screening the obtained candidate upper limb color set, candidate allergy feature set, and candidate scar feature set respectively to obtain the upper limb color, allergy feature, and scar feature included in the upper limb status information; The upper limb arm circumference included in the upper limb status information is determined according to the regional cross-sectional perimeter of the upper limb three-dimensional model in the at least three upper limb measurement areas.

7. A lymphedema identification device for breast cancer surgery, characterized in that: include: a construction unit configured to construct a three-dimensional upper limb model based on the acquired upper limb image sequence, wherein the three-dimensional upper limb model is a local three-dimensional model corresponding to at least three upper limb measurement areas acquired from multiple angles for the object to be identified; an upper limb state recognition unit configured to perform upper limb state recognition based on the upper limb image sequence and the upper limb three-dimensional model to generate upper limb state information, wherein the upper limb state information includes: upper limb dimensional features and upper limb skin features, the upper limb dimensional features include: upper limb arm circumference, and the upper limb skin features include: upper limb color, allergy features, and scar features; a first generating unit configured to generate lymphedema identification information based on the upper limb status information, the collected chief complaint information, breast cancer surgical treatment information corresponding to the subject to be identified, and a pre-trained lymphedema identification model, wherein the chief complaint information is a description of the upper limb status provided by the subject to be identified, and the breast cancer surgical treatment information includes: surgical site, surgical incision type, and radiotherapy information; wherein the lymphedema identification model includes: a chief complaint information encoder, a chief complaint information decoder, an input layer, K hidden layers, and an output layer, wherein the input layer includes: 8 neurons, and the output layer includes: 2 neurons; the lymphedema identification model and the upper limb skin feature extraction model are trained as a whole through a supervised training method, wherein the upper limb skin feature extraction model is used to extract upper limb skin features; The second generating unit is configured to generate rehabilitation advice information corresponding to the subject to be identified based on the lymphedema identification information in response to the lymphedema identification information indicating that the subject to be identified has lymphedema, The lymphedema identification information is generated based on the upper limb status information, the collected chief complaint information, the breast cancer treatment information corresponding to the subject to be identified, and the pre-trained lymphedema identification model, including: Performing text encoding on the chief complaint information by the chief complaint information encoder to obtain chief complaint information features; Decoding the features of the chief complaint information by the chief complaint information decoder to obtain a chief complaint description feature; According to the chief complaint description features, the upper limb status information and the breast cancer surgery treatment information, lymphedema identification information is generated through the input layer, the K hidden layers and the output layer.

8. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 6.

9. A computer-readable medium, characterized in that A computer program is stored thereon, wherein when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images